The same person, two calculators, hundreds of calories apart
Type your age, height, weight and activity level into one calorie calculator and it says 2,482. Type exactly the same numbers into another and it says 2,798. Neither one is broken. They are doing different arithmetic and neither of them tells you which arithmetic it picked.
That matters, because a 300-calorie error is not a rounding difference. It is the entire size of a moderate daily deficit. Aim for maintenance using the high number and you gain slowly while believing you are holding steady. Aim for a deficit using the same high number and you lose far less than you planned, then conclude your metabolism is broken.
This article takes the gap apart. There are exactly two places it comes from, and one of them is much bigger than people assume. Our own daily calorie calculator shows both steps on screen for this reason — a number you cannot inspect is a number you cannot correct.
Step one: your BMR, and the two equations that disagree
Basal metabolic rate is the energy your body uses at complete rest: breathing, circulation, keeping tissue alive. It is the floor that everything else is stacked on top of. Nobody measures your BMR directly — that requires indirect calorimetry in a lab — so calculators estimate it from height, weight, age and sex using a published equation.
Two equations dominate. The Harris-Benedict equation dates from 1919 and was revised by Roza and Shizgal in 1984; the revised version is the one still in circulation. The Mifflin-St Jeor equation was published in 1990 from a larger and more representative sample, and it is the one most clinical dietitians reach for today. Written out, for weight in kilograms, height in centimetres and age in years:
- Mifflin-St Jeor, men: 10 × weight + 6.25 × height − 5 × age + 5
- Mifflin-St Jeor, women: 10 × weight + 6.25 × height − 5 × age − 161
- Harris-Benedict (revised), men: 88.362 + 13.397 × weight + 4.799 × height − 5.677 × age
- Harris-Benedict (revised), women: 447.593 + 9.247 × weight + 3.098 × height − 4.330 × age
Take a man of 35, 80 kg, 175 cm. Mifflin-St Jeor returns 1,724 kcal. Revised Harris-Benedict returns 1,801 kcal. The gap is 77 kcal per day — real, but small. Over a week it is 539 kcal, about a seventh of a pound of body fat at the conventional 3,500 kcal per pound. If the equation were the whole story, calculators would agree within about 5% and nobody would notice.
They do not agree within 5%. So the equation is not the whole story.
Step two: the activity multiplier, where the real gap opens
BMR is then multiplied by an activity factor to reach total daily energy expenditure — the number you actually eat to. Almost every calculator on the internet uses the same five-rung ladder:
| Label | Multiplier | Usually described as |
|---|---|---|
| Sedentary | 1.2 | Desk job, little deliberate exercise |
| Lightly active | 1.375 | Light exercise 1–3 days a week |
| Moderately active | 1.55 | Moderate exercise 3–5 days a week |
| Very active | 1.725 | Hard exercise 6–7 days a week |
| Extra active | 1.9 | Physical job, or twice-daily training |
Now look at what one rung costs. For a man of 45, 90 kg, 180 cm, Mifflin-St Jeor gives a BMR of 1,805 kcal. At "lightly active" his estimate is 2,482 kcal. At "moderately active" it is 2,798 kcal. Moving one dropdown position added 316 kcal — more than double the 134 kcal that switching between the two BMR equations costs him.
This is the honest answer to why calculators disagree. The equations differ modestly. The multiplier you select differs enormously, and it is the one input that is pure self-assessment. There is no measurement behind it, only your opinion of your own week.
The systematic error runs in one direction. Someone with a desk job who trains for forty minutes four times a week reads "moderate exercise 3–5 days a week" and selects moderately active — but that ladder was built to describe the whole day, not the gym portion of it. Forty minutes of training is roughly 300–400 kcal. The rest of that person's day is sedentary. Non-exercise activity — walking, standing, fidgeting, housework — moves far more energy across a week than most training sessions do, and a desk job suppresses exactly that.
Two examples worked all the way through
Same inputs, both equations, so you can see where each number comes from.
Woman, 30 years, 65 kg, 165 cm, moderately active (× 1.55)
| Equation | BMR | TDEE at × 1.55 |
|---|---|---|
| Mifflin-St Jeor | 1,370 kcal | 2,124 kcal |
| Harris-Benedict (revised) | 1,430 kcal | 2,216 kcal |
| Difference | 60 kcal | 92 kcal |
Man, 45 years, 90 kg, 180 cm, lightly active (× 1.375)
| Equation | BMR | TDEE at × 1.375 |
|---|---|---|
| Mifflin-St Jeor | 1,805 kcal | 2,482 kcal |
| Harris-Benedict (revised) | 1,902 kcal | 2,616 kcal |
| Difference | 97 kcal | 134 kcal |
Follow the second man one step further, because this is where an invisible error becomes a visible failure. He wants a 500 kcal deficit. He uses the Harris-Benedict figure, subtracts 500, and eats 2,116 kcal. If his true maintenance is nearer the Mifflin-St Jeor estimate of 2,482, his actual deficit is 366 kcal — about 27% smaller than the one he planned. He will lose weight, just noticeably slower than the arithmetic on his notes app promised, and he will have no idea why.
What this cannot tell you
Every number above is a population estimate applied to an individual, and it is worth being precise about what that does and does not buy you.
- It cannot tell you your actual metabolic rate. Prediction equations are validated against lab measurement at the group level. For any single person the error can run well beyond 10% in either direction, and the equation cannot tell you which direction yours runs.
- It cannot see your body composition. Height and weight cannot distinguish muscle from fat, and muscle is more metabolically active. Two people with identical stats and different builds genuinely differ, and neither equation above has any input that would notice.
- It cannot account for what you do between workouts. Spontaneous movement varies hugely between people and is not something you can reliably self-report on a five-item scale.
- It cannot stay accurate as you change. Lose 10 kg and your BMR falls; add muscle and it rises. A number calculated once is a snapshot, not a setting.
- It is not a target handed down by anyone who knows your circumstances. It is arithmetic on four inputs.
How to pick one number and correct it from your own data
The way out is to stop shopping for a calculator that gives you the number you like, and start treating the first number as a hypothesis you test.
- Pick Mifflin-St Jeor and stay with it. It performs better than Harris-Benedict across most modern adult samples, and consistency matters more than the last 50 calories. If a calculator does not say which equation it uses, that is a reason to distrust the output, not a detail.
- Choose the activity level below the one you want to choose. This is the single highest-value correction on the page. If you sit for work, start at sedentary or lightly active even if you train — the ladder describes your day, not your gym habit.
- Hold that intake for two full weeks. One week is too short: water shifts from salt, carbohydrate and hormonal cycles routinely move the scale by more than the signal you are looking for.
- Weigh yourself daily and use the weekly average. Compare average against average, never one morning against another.
- Adjust in 100–200 kcal steps. Weight flat and you wanted flat: the estimate fits, stop. Falling faster than you intended: add 100–200. Not moving when you expected it to: subtract 100–200. Then hold another two weeks.
- Keep everything else still while you test. Starting a new training block in week one means you no longer know which change moved the scale.
After one or two cycles you have something no equation can give you: a maintenance figure derived from your own body rather than from a regression on somebody else's. From then on the calculator is only a sanity check for when your weight or routine changes substantially.
Frequently asked questions
Which calorie calculator is the most accurate?
The question is slightly wrong. Accuracy comes from the equation plus the activity multiplier plus your honesty about the second one, and the multiplier dominates. Prefer a calculator that states its equation and shows the multiplier it applied. Mifflin-St Jeor is the better default, but a stated-method calculator you then correct against two weeks of weight data beats any "most accurate" claim.
How far apart do calculators actually get?
Holding the activity level fixed, swapping between Mifflin-St Jeor and revised Harris-Benedict moved our two worked examples by 92 and 134 kcal per day. Moving one rung on the activity ladder moved the second example by 316 kcal. Combine an unlucky equation with an optimistic multiplier and 400 kcal of disagreement between two sites is entirely ordinary.
Should I use Harris-Benedict or Mifflin-St Jeor?
Mifflin-St Jeor, for most adults. It was derived from a larger and more modern sample and tends to sit closer to measured resting expenditure. Revised Harris-Benedict generally reads higher, which is the less forgiving direction to be wrong in if you are trying to lose weight. Either way, what makes the number usable is the correction step, not the formula.
How often should I recalculate?
When something material changes: body weight by roughly 5 kg or more, or a genuine shift in your routine such as a new job or a new training schedule. Recalculating weekly is noise, because the inputs barely move and the output follows them mechanically. Your own weight trend is the more informative signal between recalculations.
If you want the two steps laid out side by side with your own numbers in them, the daily calorie calculator shows the equation and the multiplier separately rather than collapsing them into a single figure. The rest of the section is in food and weight tools.
HealthGuideTip provides general health information for educational purposes only and is not medical advice. Always consult a qualified healthcare professional about your own health, and before changing any treatment, medication, or lifestyle.