Tools

TDEE calculator: 3 formulas and the real error margin

Every TDEE calculator hands you one confident number, and for one person in four that number is wrong by tens of per cent. So we give a band.

Nagerekend door Bodi · Laatst nagerekend 4 augustus 2026 · 7 bronnen

Calculate your TDEE

This calculator needs JavaScript. With it switched off, the formulas further down this page still stand. They are meant to be redone on a pocket calculator, and that is deliberate.

The three formulas are written out below, with their source, so you can redo them on a pocket calculator.

Short answer

What is TDEE?
Total Daily Energy Expenditure. Everything you burn over 24 hours, in kilocalories: what you burn at rest plus everything on top of it. Also called your maintenance level.

How do you calculate TDEE?
BMR times PAL. First your expenditure at rest, from weight, height, age and sex. Then times your PAL value: your total expenditure divided by your expenditure at rest, measured with doubly labelled water.1

What is the difference between TDEE and BMR?
BMR is what you burn lying still, TDEE is what you burn over a day. In an average adult BMR is roughly 60 per cent of TDEE. A PAL of 1.63 means 63 per cent on top of your BMR.1

How accurate is a TDEE calculation?
Within 10 per cent for three in four people, and not for the fourth. SACN put the formula next to measurements in 767 people: 73 per cent fell within 10 per cent, and for the rest the error ran to minus 33 and plus 35 per cent.1 Hence a band above, and not a single figure.

How we count on the rest of this site

What is TDEE?

TDEE stands for Total Daily Energy Expenditure: your total expenditure over 24 hours, in kilocalories. It is built from three parts.

The largest part goes on doing nothing. Your heart, kidneys, liver and brain keep running while you lie still, and that is your basal metabolic rate, or BMR. On top of that comes everything you move, from stairs to training. And third, processing your food costs kilocalories in itself, roughly a tenth of what you take in.

Those three together are your TDEE. You also hear it called your maintenance level: the point where what you take in and what you burn cancel out.

How do you calculate your TDEE?

In two steps, and every calculator does it this way.

Step 1. Your BMR. A formula estimates your expenditure at rest from weight, height, age and sex. There are several and they give different answers; the three we use are below.

Step 2. Times your PAL value. PAL stands for physical activity level and is your total expenditure divided by your expenditure at rest. At the population average of 1.63 you burn 63 per cent on top of your BMR.1

TDEE = BMR × PAL. That is all. The arithmetic is trivial; the problem sits in both numbers.

TDEE and BMR: what is the difference?

BMR is what you burn lying awake in bed doing nothing: your basal metabolic rate. TDEE is what you burn on an ordinary day, so BMR plus everything on top.

In an average adult BMR is roughly 60 per cent of TDEE, since 1 divided by 1.63 is 0.61. Someone seated all day comes closer to 72 per cent; heavy physical work drops it to about 50 per cent.1

You will also come across RMR and REE, for resting metabolic rate and resting energy expenditure. The sources under this page use those terms interchangeably for the same kind of measurement: SACN writes BMR,1 Mifflin and Frankenfield write resting energy expenditure.34 We stick to BMR. Exactly how much the measuring conditions differ is something we have not verified, so it is not stated here.

The three formulas, so you can redo them yourself

All three are published and can be redone on a pocket calculator. They estimate your BMR. That figure then goes times your PAL, and only then do you have a TDEE.

Henry, 2005

Derived from a database of 10,552 measured values, collected from 174 publications between 1914 and 2001.2 Weight in kilograms, height in metres, result in megajoules per day. One megajoule is 239 kcal.

Who BMR (MJ/day)
Male, 18 to 30 0.0600 × weight + 1.31 × height + 0.473
Male, 30 to 60 0.0476 × weight + 2.26 × height − 0.574
Male, over 60 0.0478 × weight + 2.26 × height − 1.070
Female, 18 to 30 0.0433 × weight + 2.57 × height − 1.180
Female, 30 to 60 0.0342 × weight + 2.10 × height − 0.0486
Female, over 60 0.0356 × weight + 1.76 × height + 0.0448

The age bands overlap on paper. Henry writes 18-30, 30-60 and over 60, so at exactly 30 there are two names. We read 30 and 60 as the upper band. For a man of 80 kilograms and 1.80 metres that is a difference of over 90 kcal a day, so it is not a detail.1

Mifflin-St Jeor, 1990

The formula you meet on most TDEE calculators. Derived from measurements in 498 healthy Americans aged 19 to 78.3 Weight in kilograms, height in centimetres, result already in kcal per day.

10 × weight + 6.25 × height − 5 × age, plus 5 if you are male and minus 161 if you are female.3

It is here because a 2005 systematic review called it the most reliable of the common formulas, with the narrowest error range. That same review adds straight away that noteworthy errors remain once you apply it to one individual.4 Harris-Benedict from 1919, which you still see a lot, came out worse.4

Cunningham, 1980

500 + 22 × fat-free mass in kilograms.7 It only produces a number once you enter a body fat percentage, because without that figure there is no fat-free mass to work with.

Why it belongs here: the other two formulas estimate your fat-free mass from weight, height, age and sex. Anyone carrying more muscle than average for their height is underestimated by that guess. Cunningham skips the guess.

In 90 recreational athletes training an average of 9.1 hours a week, Cunningham came out best of the formulas tested: 84.9 per cent of men and 78.4 per cent of women within 10 per cent of the measurement.5 Note the condition: in that study fat-free mass was determined in a measuring chamber. Enter an estimated percentage and the result is exactly as good as that estimate.

What is a PAL value, and which one fits you?

PAL is not a factor somebody invented. It is a measured ratio: total expenditure over a day, divided by expenditure at rest. Measured with doubly labelled water, a method where you drink labelled water and read off from what you excrete how much you have burned.

The awkward part is that a lifestyle does not predict a PAL value with any certainty. SACN says so itself. What can be priced reliably is the cost of a described activity.1 Hence the menu above talks in hours and sessions per week, and not in labels like “lightly active”.

The reference points SACN gives for that:

Adds What you do for it
+ 0.15 30 minutes of moderate activity on five or more days a week
+ 0.2 an hour of brisk walking a day, 6 to 7.5 km per hour
+ 0.3 five one-hour sessions a week, for instance running at 9 km per hour
+ 0.4 an hour of running at 9 km per hour every day
+ 0.6 daily intensive training at competitive level

The values in the menu come from the same source. SACN uses 1.49 for less active than average, 1.63 for average and 1.78 for more active than average.1 FAO, WHO and UNU split the whole range in three: 1.40 to 1.69 for sedentary or light, 1.70 to 1.99 for active, 2.00 to 2.40 for vigorous. Above 2.40 almost nobody keeps it up for long.6

Why PAL 1.2 for a desk job is too low

Most calculators offer 1.2 for a desk job. That figure comes from the American DRI report.

SACN has done the arithmetic and lands somewhere else. Someone who only gets up, showers and dresses is already at 1.35 to 1.4. In older people living at home, 1.38 was the average measured. For people who are not confined to a bed, 1.38 appears to be the floor, and SACN states in as many words that the upper limit of 1.4 from the American report is therefore too low.1

Anyone offering 1.2 is offering a value below the measured floor. The lowest option here is 1.38.

How accurate is a TDEE calculation?

A calculator saying 2,340 kcal makes two claims. The first is that your expenditure can be derived from four measurements. The second is that the final 40 means something.

The first claim is reasonable. The second is not.

SACN put the formula it uses itself next to real measurements in 767 people. Of those, 73 per cent fell within 10 per cent of the measurement. For the rest the error ran to minus 33 and plus 35 per cent.1 So one in four people gets a figure from a formula that is tens of per cent off, and nobody can tell from the figure whether they are that fourth person.

Hence three formulas side by side here rather than one, and hence the rounding to tens. The difference between the formulas is not noise we should have cleaned up. It is the result.

Why you do not add the 10 per cent for digestion separately

Processing food costs kilocalories in itself, roughly 10 per cent of what you take in. Many calculators therefore do BMR times PAL first, and then add another 10 per cent.

That is double counting. PAL values are derived from measured total expenditure, and digestion is already inside that total. SACN even writes the lowest PAL value out in full as 1 + 0.1 for processing food + 0.29 for the acts of an ordinary day, 1.39 together.1 So that 0.1 is already in there. Add it again and you land about a tenth too high.

What this figure is not

It is not a measurement. It is the output of formulas derived on groups, and the FAO, WHO and UNU report states itself that this approach does not allow universal application to an individual.6 If you really want to know, there is indirect calorimetry, and that is a measurement in a lab.

It is also not dietary advice. What you do with this figure depends on more than four measurements, and if in doubt a doctor or dietitian is the person to ask.

Sources

Open a source and you see exactly what was used from that study, which sentence on this page it supports, and what it explicitly does not support.

  1. 1Scientific Advisory Committee on Nutrition. Dietary Reference Values for Energy. London: TSO, 2011. Crown copyright, Open Government Licence.advisory report of a government committee · full text read, 3.1 MB PDF retrieved and extracted with pdftotext · read 2026-08-03
    Design
    Advisory report of the UK government's scientific advisory committee on nutrition. Not an experiment of its own: a revision of the reference values for energy, built from total expenditure measured with doubly labelled water, with an appendix that tabulates the prediction equations used in full.
    Studied in
    Not applicable as a sample of its own. The report calculates for the UK population and, for testing the equations, relies on the American DRI doubly labelled water dataset: n=767, aged 20 to 96, BMI 18.5 to 62, 334 men and 433 women.
    Compared with
    The Henry (2005) equations on weight, those of Henry on weight and height, and those of Schofield (1985) as adopted by FAO/WHO/UNU.
    Dose and duration
    Not applicable. This is not a dosing study.
    What was measured
    Predicted resting metabolic rate against measured resting metabolic rate: the percentage of people falling within 10 per cent of the measurement, the root mean squared error, and the smallest and largest percentage deviation. Also PAL values, derived from total expenditure measured with doubly labelled water divided by resting metabolic rate.
    What was found
    Table 18 gives the Henry equations on weight and height, by sex and age band, in MJ per day and in kcal per day. For men 18-30: 0.0600 x weight + 1.31 x height + 0.473 MJ/day; 30-60: 0.0476 x weight + 2.26 x height - 0.574; over 60: 0.0478 x weight + 2.26 x height - 1.070. For women 18-30: 0.0433 x weight + 2.57 x height - 1.180; 30-60: 0.0342 x weight + 2.10 x height - 0.0486; over 60: 0.0356 x weight + 1.76 x height + 0.0448. Weight in kg, height in metres. Table 20 gives the validation: Henry on weight and height reached 73 per cent within 10 per cent of the measurement, with a root mean squared error of 114 kcal per day and a per-person deviation from minus 33 to plus 35 per cent. Henry on weight alone reached 70 per cent, Schofield 69 per cent. Paragraph 180: average total expenditure sits at 1.63 times resting metabolic rate, the spread in the population runs from about 1.38 to 2.5, and 1.49 and 1.78 are appropriate for those less or more active than average. Paragraph 290: 1.38 appears to be the floor for people living at home, so the upper limit of 1.4 from the American DRI report is too low. Paragraph 289 writes the DRI report's sedentary PAL out as 1 + 0.1 (processing food) + 0.29 (acts of daily living) = 1.39. Summary S47 states that a described lifestyle does not predict a PAL value with any certainty, but that the additional cost of a described activity can be predicted with reasonable confidence. Table 13 gives those costs: plus 0.15 for 30 minutes of moderate activity on five or more days a week, plus 0.2 for an hour of brisk walking a day (6 to 7.5 km per hour), plus 0.3 for five one-hour sessions a week such as running at 9 km per hour, plus 0.4 for that same hour every day, and plus 0.6 for an intensive training programme at competitive level. Under Prediction of the BMR it states that the accuracy of the Henry equations within the current UK population still has to be established.
    Supports on this page
    The six Henry coefficients the calculator fills in, with weight in kg and height in metres. The validation the uncertainty band rests on: 73 per cent within 10 per cent, and a per-person deviation from minus 33 to plus 35 per cent. The three PAL values in the menu: 1.49 for less active than average, 1.63 for average, 1.78 for more active than average. The floor of 1.38 for people living at home, and that 1.2 as a value for a desk job sits below it. The reason we add no separate allowance for processing food: that 0.1 is already inside the PAL value, since SACN writes the sedentary PAL out itself as 1 + 0.1 + 0.29. The whole table of what a portion of movement costs in PAL, which sits behind the (i) at the menu and carries the descriptions at 1.49, 1.63, 1.78 and 2.00. And the statement that a lifestyle does not predict a PAL while the cost of a described activity can be priced. Further derived from those same PAL values by dividing 1 by them: that BMR is roughly 60 per cent of TDEE in an average adult (1 divided by 1.63 is 0.61), about 72 per cent in someone seated nearly all day (1 divided by 1.38), and about 50 per cent with heavy work (1 divided by 2.00). Also: that SACN uses the term BMR where Mifflin and Frankenfield write resting energy expenditure.
    Explicitly does not support
    This report says nothing about a nutrient, nothing about a supplement and nothing about what anyone should eat. It is a calculation framework for population groups, and it writes itself that the accuracy of the Henry equations in the current UK population has not been established. It is moreover British and not Dutch: it contains no measurement on a Dutch sample. And it carries no statement about athletes, since the validation in table 20 was done on a general adult dataset.
    Interests and funding
    Government committee. The report is Crown copyright under the Open Government Licence. Membership and acknowledgements are in the report; no commercial funder is named.

    https://assets.publishing.service.gov.uk/media/5a7edb37ed915d74e33f2d8f/SACN_Dietary_Reference_Values_for_Energy.pdf

  2. 2Henry CJK. Basal metabolic rate studies in humans: measurement and development of new equations. Public Health Nutrition 2005;8(7A):1133-1152.primary study, reanalysis of a database · not read; only the reference and scope verified, coefficients taken from source 1 · read 2026-08-03
    Design
    Reanalysis of a collected database of measured resting metabolic rate, from which new prediction equations on weight and on weight plus height were derived. Known as the Oxford equations.
    Studied in
    Data on 13,910 men, women and children from 174 publications between 1914 and 2001. After cleaning, the equations were derived from 10,552 values.
    Compared with
    The Schofield (1985) equations, which FAO/WHO/UNU used until then.
    Dose and duration
    Not applicable. This is not a dosing study.
    What was measured
    Predicted resting metabolic rate in MJ per day, from weight, height, age and sex.
    What was found
    A series of new equations was derived, the Oxford equations, differing by age band and sex. They give lower values than the Schofield equations in men aged 18-30 and 30-60 and in all women over 18.
    Supports on this page
    The origin of the formula this calculator puts first, and the reason it is called Henry. Also: that these equations were derived from a database of measured values spanning more than eighty years of literature, and not from a single sample.
    Explicitly does not support
    This source does NOT carry the coefficients in our code, because we have not read this article ourselves. Those figures come from table 18 of source 1, where they appear in full attributed to Henry, 2005. Were table 18 to contain a transcription error, that error would be in our work too. The source also says nothing about total expenditure, since it covers resting metabolic rate only.
    Interests and funding
    Not checked. We have only seen the reference and the summary figures, not the declaration of interests.

    https://doi.org/10.1079/PHN2005801 · DOI 10.1079/PHN2005801 · PMID 16277825

  3. 3Mifflin MD, St Jeor ST, Hill LA, Scott BJ, Daugherty SA, Koh YO. A new predictive equation for resting energy expenditure in healthy individuals. American Journal of Clinical Nutrition 1990;51(2):241-247.primary study, human, derivation of an equation · abstract only · read 2026-08-03
    Design
    Measurement of resting metabolic rate by indirect calorimetry in healthy adults, after which a prediction equation from weight, height, age and sex was derived by regression.
    Studied in
    498 healthy participants: 247 women and 251 men, aged 19 to 78, mean 45. 264 of normal weight and 234 with obesity.
    Compared with
    Earlier prediction equations, including Harris and Benedict from 1919.
    Dose and duration
    Not applicable. This is not a dosing study.
    What was measured
    Resting metabolic rate in kcal per day, measured by indirect calorimetry.
    What was found
    The derived equation reads, in kcal per day: 10 x weight in kg + 6.25 x height in cm - 5 x age in years, plus 5 for men and minus 161 for women.
    Supports on this page
    The second formula in the calculator, exactly as it stands here. And the sample it rests on: 498 healthy American adults aged 19 to 78.
    Explicitly does not support
    This source says nothing about total expenditure: only resting metabolic rate was measured, not what someone burns over a day. It also says nothing about athletes, since they are not a separate group in this sample, and nothing about how well the formula performs for an individual. That last point is in source 4.
    Interests and funding
    Not checked. We read only the abstract and did not separately look up the 1990 declaration of interests.

    https://doi.org/10.1093/ajcn/51.2.241 · DOI 10.1093/ajcn/51.2.241 · PMID 2305711

  4. 4Frankenfield D, Roth-Yousey L, Compher C. Comparison of predictive equations for resting metabolic rate in healthy nonobese and obese adults: a systematic review. Journal of the American Dietetic Association 2005;105(5):775-789.systematic review · abstract only · read 2026-08-03
    Design
    Systematic review within the evidence analysis programme of the American Dietetic Association. The four prediction equations most used in practice were placed next to measured values.
    Studied in
    Healthy adults with and without obesity, from the studies meeting the inclusion criteria. We have not been able to establish the exact number of studies and participants, as it is not stated in the abstract.
    Compared with
    Harris-Benedict, Mifflin-St Jeor, Owen, and the WHO/FAO/UNU equations.
    Dose and duration
    Not applicable. This is not a dosing study.
    What was measured
    The proportion of people whose predicted resting metabolic rate fell within 10 per cent of the measured value, and the width of the error range.
    What was found
    Literally from the abstract: the Mifflin-St Jeor equation was the most reliable, predicting resting metabolic rate within 10 per cent of measured in more people with and without obesity than any other equation, and it also had the narrowest error range. The abstract notes that noteworthy errors and limitations remain when the equation is applied to an individual.
    Supports on this page
    The reason Mifflin-St Jeor is in this calculator rather than Harris-Benedict: of the common formulas, this one came out best in validation. And the caveat the review attaches itself: noteworthy errors remain for an individual.
    Explicitly does not support
    The figures 82 per cent for Mifflin-St Jeor and 68 per cent for Harris-Benedict circulate widely and are attributed to this review. We did NOT find them in the abstract and the full text sits behind a paywall. Those two figures therefore do not appear on our page. This source further says nothing about athletes and nothing about European samples.
    Interests and funding
    The review was carried out within the evidence analysis programme of the American Dietetic Association, a professional body. No declaration of interests appears in the abstract.

    https://doi.org/10.1016/j.jada.2005.02.005 · DOI 10.1016/j.jada.2005.02.005 · PMID 15883556

  5. 5ten Haaf T, Weijs PJM. Resting energy expenditure prediction in recreational athletes of 18-35 years: confirmation of Cunningham equation and an improved weight-based alternative. PLoS ONE 2014;9(10):e108460.primary study, human, comparison of equations · full text read, open access via PubMed Central · read 2026-08-03
    Design
    Resting metabolic rate measured by indirect calorimetry and fat-free mass by air displacement plethysmography, a chamber that determines the volume of the body. Existing prediction equations were then applied to those measurements and a new equation was derived.
    Studied in
    90 recreational athletes: 53 men and 37 women, mean age 23.2, all between 18 and 35. They trained an average of 9.1 hours a week, spread over an average of 5 sessions. Dutch research, carried out from Amsterdam.
    Compared with
    Among others Cunningham, Harris-Benedict, Mifflin-St Jeor and De Lorenzo, plus the two equations the authors derive themselves.
    Dose and duration
    Not applicable. This is not a dosing study.
    What was measured
    The proportion of participants whose prediction fell within 10 per cent of measured resting metabolic rate, plus the root mean squared error.
    What was found
    The Cunningham equation came out best: 84.9 per cent of men and 78.4 per cent of women fell within 10 per cent of the measurement. Their own weight-based equation reached 83.0 and 75.7 per cent, the fat-free-mass one 83.0 and 72.3 per cent, and De Lorenzo 77.4 and 59.5 per cent. Measured resting metabolic rate averaged 7.68 MJ per day, or 1837 kcal. Table 4 gives the compared equations in full, including Cunningham as 22 x fat-free mass + 500 kcal per day. The authors name as limitations that their participants avoided training for 12 hours before measurement where guidelines hold 14, that fasting was 4 hours where some guidelines ask 5, and that the new equations still need validation in other cohorts.
    Supports on this page
    The reason the calculator offers a third formula as soon as someone enters a body fat percentage, and that it is Cunningham: in recreational athletes it was the most accurate of the equations tested, with 84.9 per cent of men and 78.4 per cent of women within 10 per cent. Also the rendering of the Cunningham equation itself, 22 x fat-free mass + 500.
    Explicitly does not support
    This source carries no statement about people outside 18 to 35 and none about people who do not exercise. It also says nothing about total expenditure: only resting measurements were taken, and the PAL multiplication does not come from this study. And it says nothing about how well Cunningham performs with an estimated rather than a measured body fat percentage; here fat-free mass was determined in a measuring chamber.
    Interests and funding
    The authors declare no competing interests and report no funding: These authors have no support or funding to report.

    https://doi.org/10.1371/journal.pone.0108460 · DOI 10.1371/journal.pone.0108460 · PMID 25275434

  6. 6FAO/WHO/UNU. Human energy requirements: report of a Joint FAO/WHO/UNU Expert Consultation, Rome, 17-24 October 2001. FAO Food and Nutrition Technical Report Series 1. Rome: Food and Agriculture Organization of the United Nations, 2004. ISBN 92-5-105212-3.report of an international expert consultation · chapter on PAL and the table of lifestyle bands read; the full report of over 100 pages not · read 2026-08-03
    Design
    Report of a joint expert consultation of FAO, WHO and UNU. Not an experiment of its own: a revision of the international reference values for energy requirements, built from total expenditure measured with doubly labelled water.
    Studied in
    Not applicable as a sample of its own. The report calculates for population groups worldwide.
    Compared with
    Earlier reference values, and its own Schofield (1985) equations, which the consultation retained after deliberation.
    Dose and duration
    Not applicable. This is not a dosing study.
    What was measured
    Total expenditure expressed as a multiple of resting metabolic rate, that is, the PAL value.
    What was found
    The PAL values adult population groups sustain over long periods run from about 1.40 to 2.40. The report divides them in three: a sedentary or light activity lifestyle 1.40 to 1.69, an active or moderately active lifestyle 1.70 to 1.99, and a vigorously active lifestyle 2.00 to 2.40, noting that values above 2.40 are difficult to maintain over a long period. On the factorial approach it notes that this does not allow universal application per individual.
    Supports on this page
    The top of the menu in the calculator: PAL 2.00 as the start of the band for heavy physical work or daily vigorous training, and that above 2.40 almost nobody keeps it up for long. And the division into three lifestyle bands into which the individual values from source 1 fall.
    Explicitly does not support
    This report says nothing about a nutrient and nothing about a supplement. It is moreover written explicitly for population groups and not for individuals: it states itself that the factorial approach does not allow universal application per individual. Anyone applying these figures to themselves is doing something the report itself does not do, and that is precisely why the calculator shows a band rather than a figure.
    Interests and funding
    Report of three United Nations organisations. No commercial funder named.

    https://www.fao.org/4/y5686e/y5686e07.htm

  7. 7Cunningham JJ. A reanalysis of the factors influencing basal metabolic rate in normal adults. American Journal of Clinical Nutrition 1980;33(11):2372-2374.primary study, reanalysis · not read; only the rendering of the equation in table 4 of source 5 · read 2026-08-03
    Design
    Reanalysis of existing measurements of resting metabolic rate, from which an equation on fat-free mass was derived instead of on body weight.
    Studied in
    Adults from the reanalysed measurements. We have not established the number or composition.
    Compared with
    Equations on body weight, height, age and sex.
    Dose and duration
    Not applicable. This is not a dosing study.
    What was measured
    Resting metabolic rate in kcal per day, predicted from fat-free mass.
    What was found
    The equation reads 500 + 22 x fat-free mass in kg, in kcal per day. That is how it appears in table 4 of source 5.
    Supports on this page
    The third formula in the calculator, which appears as soon as someone enters a body fat percentage.
    Explicitly does not support
    This source has NOT been read by us. The equation was taken from table 4 of source 5, where it appears with attribution. We can therefore say nothing about the sample it rests on, the spread around it, or who it is intended for. What we do know about its accuracy we know from source 5 and not from here.
    Interests and funding
    Not checked, since the text has not been read.

    https://doi.org/10.1093/ajcn/33.11.2372 · DOI 10.1093/ajcn/33.11.2372