How to Find Your Real Maintenance Calories in 3 Weeks
A calculator gives you a maintenance estimate. How to turn it into a real number in three weeks, using weekly bodyweight averages and an honest food log.

Your real maintenance calories are the daily intake at which your weekly average bodyweight holds steady. A calculator gets you within about 10 percent. To find the real number, eat the estimate for two to three weeks, weigh yourself every morning, compare the weekly averages, and adjust by 100 to 150 calories at a time.
That is the whole method. The rest of this article is how to do each step without fooling yourself, because most people who "tried maintenance and it did not work" changed the target after five days of noisy scale readings.
Why the calculator number is only a starting point
Every calorie calculator, including ours, does two things. It estimates your resting metabolism from weight, height, age and sex, then multiplies that by an activity factor. Both steps carry error.
The resting part is decent. The Mifflin-St Jeor equation was built from 498 people measured with indirect calorimetry (Mifflin, 1990). A systematic review of the four common equations found it predicted resting metabolic rate within 10 percent of the measured value in more people than any other, with the narrowest error range (Frankenfield, 2005). Within 10 percent is the good news. The bad news is that 10 percent of a 1,800 calorie resting rate is 180 calories, and that is before the activity multiplier.
The activity multiplier is where the big errors live. "Moderately active, 1.55" is a guess about how much you walk, fidget, stand and train. Two people with the same stats and the same gym schedule can differ by 500 calories a day because one has a desk job and drives, and the other is on their feet all day. No equation sees that. Your scale does.
So treat the calculator output as a hypothesis. A good one, but a hypothesis.
The three-week method, step by step
1. Get the estimate
Run the calorie calculator with honest inputs. For the activity level, if you lift four times a week and sit the rest of the day, pick moderate, not very active. Lifting burns less than people think. A hard hour with the barbell is usually 250 to 400 calories, and that is already inside the moderate multiplier.
Write the number down. Say it gives you 2,700.
2. Eat that number for 14 to 21 days
Log everything. Weigh food in grams for the first two weeks, especially oils, nut butter, rice and anything you pour. Eyeballed portions drift upward by 10 to 20 percent, which is enough to turn a maintenance test into a slow bulk.
Keep the rest of your life the same. Same training program, same rough step count, same sleep. If you start a new program, add cardio and start the maintenance test in the same week, you will not know which change moved the scale.
3. Weigh yourself every morning
After the bathroom, before food or water, same scale, same spot on the floor. Write the number down every day. You are not going to react to any single reading. You are collecting seven of them to average.
4. Compare weekly averages, skipping week one
Week one is noise. If you were dieting before, your glycogen refills and you gain a kilo of water in four days. If you were overeating before, you drop some. Neither is fat.
Take the seven-day average of week two and week three and compare them. Three outcomes:
- The averages are within about 0.3 kg of each other. You are at maintenance. Done.
- Week three is lower than week two by 0.3 kg or more. You are under maintenance.
- Week three is higher by 0.3 kg or more. You are over maintenance.
5. Turn the trend into calories
A rough working rule: one kilogram of bodyweight change is about 7,700 calories of energy balance. So if your weekly average dropped 0.3 kg in a week, you were under by roughly 7,700 × 0.3 ÷ 7, which is about 330 calories a day.
That rule is not exact. Hall's group at the NIH showed the body's response to a change in intake is slow and non-linear, with a half-life of around a year (Hall, 2011), and Thomas and colleagues showed the classic "3,500 calories per pound" version overestimates fat loss the longer a diet runs (Thomas, 2013). Over a two-week window, though, it is close enough to point you in the right direction.
6. Adjust in small steps and repeat
If you were under by 300, add 150 and run another two weeks. If you were over, subtract 150. Small steps, because a 400 calorie swing overshoots and you spend another month bouncing around the answer.
Most people land on their number inside four to five weeks total. Then it is yours until your routine changes.
A worked example
Mark is 82 kg, 178 cm, 29 years old, lifts four days a week and works at a desk.
Mifflin-St Jeor: 10 × 82 + 6.25 × 178 − 5 × 29 + 5 = 1,793 calories at rest. Times 1.55 for moderate activity gives 2,778. He rounds to 2,780 and eats it, weighing everything.
His weekly averages:
| Week | Average bodyweight | Change |
|---|---|---|
| 1 | 82.6 kg | +0.6 kg (water, more carbs than before) |
| 2 | 82.8 kg | +0.2 kg |
| 3 | 83.0 kg | +0.2 kg |
Week two to week three is +0.2 kg. That is roughly 7,700 × 0.2 ÷ 7, about 220 calories a day over. His real maintenance is closer to 2,560 than 2,780. The desk job is the likely reason: his multiplier was more like 1.42 than 1.55.
He drops to 2,600 and runs two more weeks. Averages: 83.0, then 82.9. Flat. His maintenance is about 2,600 calories, and now his cut or his lean bulk starts from a number that is real.
Notice what he did not do. He did not panic in week one at +0.6 kg. He did not decide the calculator was broken. He changed one thing by 180 calories and waited.
A shortcut if you already have the data
If you have logged your food for a few months and your bodyweight has been stable across that time, you do not need the three-week test. Your average daily intake over any stable four-week stretch is your maintenance. Add up the calories, divide by the days, and that number is worth more than any equation, because it was measured on you.
The catch is the word stable. If the average drifted up or down by more than half a kilo across the month, the intake was not maintenance, it was a slow surplus or deficit. Correct for it: roughly 7,700 calories per kilogram of drift, divided over the days. A month at 2,900 a day with 0.5 kg gained is a maintenance of about 2,770.
Mistakes that wreck the estimate
Picking an activity level you wish you had. "Very active, 1.725" is for people who do physical work all day and train on top. Four gym sessions and a desk is moderate at best. This single choice is the most common reason a calculator overshoots by 400 calories.
Changing three things at once. New program, new step goal and a new calorie target in the same week. When the scale moves, you cannot say why.
Weighing at random times. A weigh-in after dinner on a high-salt day is two kilos off a morning reading. Same time, same conditions, every day.
Guessing the calorie-dense stuff. Oil, butter, peanut butter, cheese, granola. A tablespoon of olive oil is 120 calories, and a "tablespoon" poured from the bottle is often two.
Letting weekends off the log. Two restaurant meals and a few drinks can add 1,500 calories to a week. Spread across seven days that is 200 a day, which looks exactly like the calculator being wrong.
Quitting after one week. Week one is water. The people who say "I tried 2,700 and gained a kilo" usually stopped there.
Trusting the watch. Fitness trackers guess energy burn from heart rate and movement. The number is a guess wearing a decimal point. Your weekly average is the measurement.
How to track this in Ultra+
The Ultra+ nutrition journal has food search and barcode scan, so the "weigh and log everything" step takes seconds per meal. Scan the packet, type the grams, done. It shows your protein, carbs and fat for the day next to your calorie target, and it logs your bodyweight in the same place, so your intake and your scale trend sit side by side.
Keep training the same while you test. Ultra+ shows your previous set and your best set on every set you log, so if strength starts sliding during the test you will see it in the numbers rather than wondering. At the end of the week, the weekly card reviews the sessions you actually did, which tells you whether the training side of the equation held still.
What to do once you know the number
Once you have a maintenance figure you trust, everything downstream gets easier. A cut becomes maintenance minus 300 to 500, and you can read how big a deficit should be if you lift for the reasoning. A lean bulk becomes maintenance plus 10 to 20 percent, covered in lean bulk calories explained.
Then recheck it. Maintenance moves when your steps, sleep or bodyweight move, and why your maintenance calories change walks through how much each one shifts the number. The habit that matters is the one you just built: log honestly, weigh daily, judge by the weekly average, and change one thing at a time.
Frequently asked questions
- How long does it take to find maintenance calories?
- Two to three weeks of eating the same target and weighing yourself every morning. The first week is mostly water moving around, so compare the average of week two with the average of week three. If they are within about 0.3 kg of each other, you have your number.
- How accurate is a maintenance calorie calculator?
- Close, not exact. The Mifflin-St Jeor equation lands within 10 percent of measured resting metabolism for most people, and the activity multiplier adds its own error. A 2,700 estimate can mean a real maintenance anywhere from about 2,450 to 2,950. That is why you test it.
- Should I weigh myself every day to find maintenance?
- Yes, but only look at the weekly average. Daily weight swings a kilo or more with water, salt, carbs and what is in your gut. Weigh after the bathroom and before food, write it down, and compare seven-day averages. One weigh-in tells you nothing.
- What if my weight goes up in the first week at maintenance?
- That is usually glycogen and water, especially if you were dieting before. A gram of stored carbohydrate holds around three grams of water. Hold the target, keep logging, and judge it from week two onward. Do not cut calories because of week one.
- Does maintenance change once I find it?
- Yes. It moves with your steps, your training volume, your sleep and your bodyweight. A number you found in a busy summer can be 200 to 300 calories too high in a desk-bound winter. Recheck it whenever your routine or your bodyweight changes by more than a few percent.
References
Every study below was opened and read before it was cited.
- A new predictive equation for resting energy expenditure in healthy individualsMifflin MD, Am J Clin Nutr, 1990
- Comparison of predictive equations for resting metabolic rate in healthy nonobese and obese adults: a systematic reviewFrankenfield D, J Am Diet Assoc, 2005
- Quantification of the effect of energy imbalance on bodyweightHall KD, Lancet, 2011
- Can a weight loss of one pound a week be achieved with a 3500-kcal deficit? Commentary on a commonly accepted ruleThomas DM, Int J Obes, 2013
Related
Written by Sebastian, founder, ultra+. Sebastian founded Ultra+ and lifts. He writes every article on this site and checks each cited study himself.

