How to Count Calories to Lose Weight: Deficit Size, Pace, and When to Stop
Eat 10–20 percent below what you burn, expect to lose 0.25–1 percent of your bodyweight a week, never let the target drop below your own resting metabolic rate, and recalculate after every 3–4 kilograms, because a lighter body burns less. That is the whole plan in one sentence. The rest is about why those numbers, why the scale slows after a few weeks when you have changed nothing, and how to decide in advance when to pause and when to stop.
What this post assumes
It assumes you already know how to count — what to log, how precise to be, how to keep the habit past week two. If not, start with our guide to counting calories and come back. It also assumes you accept that no number here is exact — labels, databases and equations all carry error, which we document separately — so you read your weight as a weekly average and your intake as an estimate.
Disclosure: Olivka is our product. It appears once, near the end, and in one FAQ answer; it is a free trial followed by a paid subscription, and nothing in the numbers below depends on using it.
One more thing this post assumes: that you want to count. If the arithmetic itself is the part you dread, there are approaches that get most of the way without it — losing weight without counting calories sets out what works and where it stops working.
How big a deficit: 10–20 percent, and not below your BMR
A deficit is the gap between what you burn in a day — your total daily energy expenditure, or TDEE — and what you eat. The calorie calculator estimates TDEE with the Mifflin-St Jeor equation from your age, sex, height, weight and activity. Take 10–20 percent off that figure and you have a target. At a 2,300 kcal expenditure that is 1,840–2,070 kcal a day.
Better to size it as a rate than as a fixed gap: the same 500 kcal is gentle at 95 kg and severe at 55 kg. Aim to lose 0.25–1 percent of your bodyweight a week: 0.5 percent is the sensible default, 1 percent the ceiling. Past 1 percent, what rises is not the rate fat comes off but the share of muscle in what you lose.
A 2011 trial in elite athletes randomised 24 of them to lose 0.7 or 1.4 percent of bodyweight a week, all lifting four times weekly. Both groups lost about 5.5 percent of their bodyweight. The slower group lost more of it as fat (31 percent of fat mass versus 21) and gained 2.1 percent lean mass; the faster group held level. Athletes are not most readers, but the direction is what generalises.
The floor is your basal metabolic rate — what your body uses at complete rest. Eating below it is not faster weight loss; it is running those functions at a discount, and the bill arrives as fatigue, worse sleep, muscle loss and, in women, often a disrupted cycle. That is why the deficit calculator refuses to return a target below your BMR: it raises the target to that level, says so, and shows the slower rate the capped plan delivers.
The popular “1,200 kcal a day” figure is not used anywhere on this site. It is a constant applied to everyone, whereas BMR is the same idea computed for the actual person: if yours comes out at 1,550, the constant would have you 350 kcal under your floor from day one.
How long it takes: the 7,700 kcal rule and where it breaks
Every deficit calculator, ours included, uses the same approximation: a kilogram of body fat holds roughly 7,700 kcal, so a 550 kcal daily deficit should cost about half a kilo a week. For the first month or two it is about right; as a forecast for six months it is wrong in two known ways.
The first is that not everything you lose is fat. A 2008 analysis in the International Journal of Obesity modelled the composition of weight loss and found the rule holds for people who begin with more than about 30 kg of fat, whose loss is almost all fat. For leaner people a larger share of each kilo is lean tissue, which stores far less energy — so they lose weight faster per calorie of deficit than the rule predicts, and less of it is the weight they wanted to lose.
The second is that your expenditure falls as you shrink. A lighter body costs less to move and to run at rest, and a sustained deficit lowers resting expenditure somewhat beyond that. A 2011 model in The Lancet that accounts for these adaptations finds that after a permanent cut in intake it takes about a year to reach half of the eventual weight change and about three years to reach 95 percent of it. The linear rule predicted roughly twice as much loss in the first year for the same 2,000 kcal-a-day cut.
From the inside: the first two weeks come off fast, partly water and stored carbohydrate rather than fat, and then the weekly number settles below what the arithmetic said. That is not a stall. It is the reason to recalculate after every 3–4 kg and to read a weekly average, weighed at the same time of day, instead of any single morning — bodyweight swings a kilo overnight on salt, glycogen, your cycle and digestion.
One reader, one TDEE, one plan
Take a 34-year-old woman, 168 cm, 78 kg, training three or four times a week. Mifflin-St Jeor gives her a BMR of about 1,500 kcal, and the moderate activity multiplier brings her TDEE to about 2,320 kcal. Every figure is rounded to 10 kcal; the equation carries roughly a 10 percent standard error.
| Rate of loss | Daily deficit | Daily target | What happens |
|---|---|---|---|
| 0.25 % a week (0.19 kg) | 210 kcal | 2,110 kcal | Barely a diet; slow, easy to hold for months |
| 0.5 % a week (0.39 kg) | 430 kcal | 1,890 kcal | The default — about 18 % below TDEE |
| 0.75 % a week (0.58 kg) | 640 kcal | 1,680 kcal | Fine for a few weeks; hunger and sleep start to cost |
| 1 % a week (0.78 kg) | 820 kcal | 1,500 kcal | Hits her BMR floor — capped here, delivers 0.75 kg a week |
Say she takes the 0.5 percent row. After roughly ten weeks she is 4 kg lighter, and if she re-runs the numbers at 74 kg her TDEE has dropped to about 2,260 kcal and the same rate now means 1,850 kcal a day. Forty calories is not dramatic; skipping the recalculation three times in a row is, because by then the plan she thinks she is on and the one she is actually on differ by a full rate band. If she wants 8 kg, the second half will take longer than the first, which is where a maintenance break belongs.
Protein is the second number
Once the calorie target is set, the one macronutrient worth tracking alongside it is protein. A 2012 meta-analysis in the American Journal of Clinical Nutrition pooled 24 randomised trials with 1,063 participants comparing higher-protein and standard-protein diets at the same energy restriction. The higher-protein groups lost 0.79 kg more weight, 0.87 kg more fat and 0.43 kg less fat-free mass, and their resting expenditure fell by about 140 kcal a day less. The authors called the benefit modest, and it is — but it points the same way on every measure that matters in a deficit.
Our macro calculator sets protein at 2 g per kilo of bodyweight when the goal is loss — about 156 g a day for the reader above, with fat floored at 0.8 g per kilo and carbohydrate taking the rest. You do not have to hit 156 exactly; you do have to notice that breakfast and snacks are where protein usually goes missing.
When to pause: the case for a maintenance break
A deficit does not have to be continuous, and there is a reasonable case that it should not be. The MATADOR trial (International Journal of Obesity, 2018) put 51 men with obesity on 16 weeks at 67 percent of their maintenance needs, either continuously or in two-week blocks alternating with two weeks at maintenance. Among completers the intermittent group lost 14.1 kg against 9.1 kg, almost all of the difference as fat, with no extra loss of fat-free mass, and six months later they were still about 8 kg further below their starting weight.
It is one trial, in men, analysed per protocol, at a deficit deeper than anything recommended above — a strong hint rather than a law. The practical version: after 8–12 weeks in a deficit, or after each 3–4 kg lost, spend one to two weeks eating at your newly recalculated maintenance. The scale rises a little in the first days — water and glycogen, not fat — and you come back with a fresh target and a rehearsal of the eating you will have to keep up once you stop.
When to stop
Decide this before you start, because the deficit itself will not tell you. Three exits are normal and one is urgent.
- You reached the weight, waist or strength-to-weight you set out for. Do not jump back to your old intake: add 100–150 kcal a day, hold for two weeks, repeat until your weekly average stops falling. That number is your real maintenance.
- Your four-week average has not moved and the counting is honest. Add daily movement, or take off 100–150 kcal, never more — or, often better, spend two weeks at maintenance and resume.
- You have been at it for four to six months. A planned month or two at maintenance, then a second round if you still want one, beats a fifth month of grinding.
- Tracking has stopped informing you and started changing how you feel about food or yourself — skipping meals with other people because they are hard to log, or eating worse rather than better. Stop, and talk to a doctor or a dietitian rather than pushing through. No target on any calculator is worth that.
If you are pregnant or breastfeeding, under 18, managing a medical condition, or taking medication that affects appetite or metabolism, set any calorie target with a clinician rather than with a calculator. The ranges in this post are for healthy adults.
The fourth case is the awkward one: the deficit is set, the logging is honest, and the scale has not moved in a month. That has its own causes, in a fairly predictable order — counting calories but not losing weight works through them.
What the day-to-day looks like
The plan above is a handful of numbers held for a few months. The work is daily: knowing by evening where you are against the target, because a deficit you cannot see is one you are guessing at.
Olivka moves that part into a chat you already have open. You send a photo of the plate, a voice note or a sentence on WhatsApp, Messenger, Telegram or Instagram; a vision model reads the plate or a nutrition label, ingredients are matched against USDA FoodData Central and Open Food Facts, and you get an itemised meal card back with Confirm and Edit. Corrections replace the earlier entry — “there was mayo in it too” updates the meal and the day. You set the daily calorie target and per-macro gram targets in the chat or on the web dashboard, and the running total is read against them all day.
What it will not do is choose the deficit for you, plan your meals, or tell you what to weigh. Those decisions stay yours.
Frequently asked questions
How many calories should I eat a day to lose weight?
Ten to twenty percent below your total daily energy expenditure, which you estimate from age, sex, height, weight and activity with the Mifflin-St Jeor equation — a free calculator does it in a minute. At a 2,300 kcal expenditure that is roughly 1,840–2,070 kcal a day. Never take the target below your own basal metabolic rate, and ignore any single number quoted as universal, including the popular 1,200 kcal: it is a constant applied to everyone, whereas your BMR is the same idea computed for you. Recalculate after every 3–4 kg lost, because a lighter body burns less.
How fast should I lose weight when counting calories?
Between 0.25 and 1 percent of your bodyweight a week, with 0.5 percent as the sensible default — about 0.4 kg a week for someone at 78 kg. Past 1 percent the extra speed comes mostly from muscle rather than fat: in a randomised trial of athletes, the group losing 0.7 percent a week lost more fat and gained lean mass, while the group at 1.4 percent lost less fat and no lean mass at all. Expect the first two weeks to run faster than the rate implies, because water and stored carbohydrate go first, and the weeks after to run a little slower than the arithmetic predicts.
Why has my weight loss slowed down even though I am eating the same?
Two reasons, both normal. Your expenditure has fallen: a lighter body costs less to move and run at rest, and a sustained deficit lowers resting expenditure somewhat beyond that. And the 7,700 kcal-per-kilo rule is a static approximation — a 2011 model in The Lancet that accounts for these adaptations finds it takes about a year to reach half of the eventual weight change from a permanent cut in intake. Recalculate your target at your new weight, read a weekly average instead of single weigh-ins, and if the average is genuinely flat for a month, add movement, trim 100–150 kcal, or take two weeks at maintenance.
Should I take breaks from a calorie deficit?
There is a reasonable case for it. In the MATADOR trial, 51 men with obesity ate at 67 percent of maintenance for 16 weeks either continuously or in two-week blocks alternating with two-week maintenance blocks; the intermittent group lost about 5 kg more, nearly all of it fat, and was still about 8 kg further below baseline six months later. It is one trial in one population, so treat it as a hint. A practical version: after 8–12 weeks, or after each 3–4 kg, spend one to two weeks at your recalculated maintenance, then resume with a fresh target.