Energy · 14 min read

TDEE Explained: Why the Number Is a Starting Point

Written by Coach Ohm · reviewed 2026-08-30

Try it: TDEE and Calorie Target Calculator — Your maintenance calories, plus a target that matches the rate of loss you can hold.

Every calorie calculator gives you a confident number. None of them measured anything. Here is what the number is built from and how much of it you should trust.

The four parts

Total daily energy expenditure is the sum of four things:

Basal metabolic rate. What you burn keeping the lights on — organs, brain, temperature. For most people this is the largest single share, typically somewhere in the region of 60 to 75 per cent of the total.
The thermic effect of food. The energy cost of digesting what you eat, commonly cited around 8 to 10 per cent. Protein costs the most to process, which is a small side benefit of a higher-protein diet rather than a strategy on its own.
Deliberate exercise. Usually far less than people assume. An hour of training is real but it is not the biggest lever.
Everything else you do. Walking, standing, fidgeting, carrying things. This varies enormously between people and, importantly, it varies within the same person from week to week.

Where the estimate goes wrong

The basal part is calculated from height, weight, age and sex, and it is reasonably reliable. The problem is the activity multiplier. Two people select "trains three times a week" — one of them also walks 12,000 steps a day and the other drives everywhere. The gap between them is several hundred calories, and the calculator cannot see it.

That is the honest limitation. A TDEE calculator narrows the range from "no idea" to "probably within a few hundred calories". It does not produce a measurement.

Walk through the arithmetic yourself

Here is the whole calculation for a concrete person, step by step, so you can see where each number comes from and where it could be wrong. Person A: 32 years old, 175 cm, 80 kg, male, desk job and a gym habit he rates as moderate — training three to five days a week.

The basal part uses the equation the site's calculator runs on. It is not a measurement; it is a calculation from your body size and age. For Person A it lands at 1,739 calories. That is what his body would burn lying perfectly still, awake, doing nothing except being alive.

Multiplying by his activity level, which the calculator sets at 1.55 for "moderate, train 3–5 days", gives a maintenance estimate of 2,695. That is the number every calculator is handing him, and it is also the number most likely to be wrong. The reason is in the next section.

Now set a target. Person A wants to lose weight at a pace he can hold, so he picks the slowest option the calculator offers: 0.25 per cent of bodyweight per week. At 80 kg that is 0.2 kg a week. Fat is stored at roughly 7,700 calories per kilogram, so 0.2 kg costs about 1,540 calories a week — 220 a day. His daily target is 2,695 minus 220, which is 2,475. If he is training with any real intent, his protein target sits at 128 to 160 grams a day — a separate number from the calorie target, and one the calorie target alone will never tell him.

Run the same steps for a different person

Person B is a 25-year-old woman, 160 cm, 65 kg, desk job, little structured exercise. Same equation, same method. Her basal figure comes out to 1,364 — noticeably lower than Person A's, and that gap is almost entirely body size, age and sex, not any quality or failure on her part. Multiply by her activity level, 1.2 for "desk job, little exercise", and her maintenance estimate is 1,637.

She wants to lose weight faster than Person A is willing to, so she picks 0.5 per cent of bodyweight a week, which at 65 kg works out to 0.33 kg a week. At 7,700 calories per kilogram of fat, that costs roughly 358 calories a day. Her target is 1,637 minus 358, which is 1,279 — comfortably above the calculator's built-in floor of 1,200 for women, so nothing gets clipped. Her protein target, if she trains seriously, sits at 104 to 130 grams a day.

Notice what actually differs between the two plans, and it is not simply "her numbers are smaller." Person A's daily deficit of 220 calories is about 8 per cent of his 2,695-calorie maintenance. Person B's daily deficit of 358 calories is about 22 per cent of her 1,637-calorie maintenance — nearly three times the relative size of his, because she chose the faster rate, not because her body is doing anything different. A faster target rate is always a proportionally bigger bite out of a smaller number, and a smaller number has less room to absorb it. If someone hands you a calorie target that was built for a bigger body at a slower rate, it is not the same plan with the numbers shrunk — it is a different plan entirely.

A third example: same equation, an older and heavier body

Person C is 45, male, 180 cm, 95 kg, and trains one to three days a week — the calculator's "light" band, multiplier 1.375. His basal figure comes out to 1,855, higher than either Person A or Person B because he is carrying more bodyweight, and his maintenance estimate is 2,551.

He also wants to move faster than Person A: 0.75 per cent of bodyweight a week, which at 95 kg is 0.71 kg a week, costing roughly 784 calories a day. His target is 2,551 minus 784, which is 1,767. His protein target, because he is still lifting through the deficit, is 152 to 190 grams a day — noticeably higher than either Person A's or Person B's, simply because he is a bigger man with more muscle to hold onto.

Person C's numbers make a point the other two do not: age is genuinely in the equation, but it is a small term next to bodyweight and activity. A 45-year-old at 95 kg training lightly still comes out with a higher basal figure than a 32-year-old at 80 kg training harder. If you are older and worried that your metabolism has quietly collapsed with age, look at your weight and your activity level before you look at the calendar. The equation already accounts for age; it is just not the part doing most of the work.

When body composition changes the picture

Mifflin-St Jeor, the equation behind all three of these numbers, works from height, weight, age and sex. It does not know how much of that weight is muscle and how much is fat, and two people who weigh the same can have genuinely different basal rates because of it. If you know your body fat percentage, the site's calculator can switch to Katch-McArdle instead, which works from lean mass rather than total bodyweight. Take someone at 80 kg carrying 18 per cent body fat: that is 65.6 kg of lean tissue, and Katch-McArdle puts their basal rate at 1,787 calories — 48 calories higher than the 1,739 Mifflin-St Jeor gives an 80 kg man by height and age alone, because a leaner 80 kg carries more of the tissue that actually burns energy. The gap is not large for most people, which is why the height-weight-age equation is the sensible default. It grows for anyone notably leaner or fatter than average at a given weight — a lean, muscular person will be under-estimated by the default equation, and a person carrying more fat at the same scale weight will be over-estimated. If your result feels consistently wrong in one direction and you know your body fat percentage, that is the first thing to check before assuming the equation itself is broken.

Where the multiplier does all the damage

Look at the number that actually moves the total. Person A's basal figure is fixed at 1,739 no matter what you do with the rest of the calculation — it does not change unless his height, weight, age or sex changes. The total is that fixed number times his multiplier. Move one row up the activity table, from "moderate, train 3–5 days" to "high, train 6–7 days", or one row down to "desk, little exercise", and his maintenance shifts by several hundred calories a day in either direction, because the multiplier is applied to the whole basal figure, not a slice of it. One decision — which row he picks — moves his entire daily number by roughly a third of his total.

That is why the multiplier, and not the equation, is where the estimate lives or dies. The equation is a sensible way to get a basal number from your size and age, and it is the part of the total that is least likely to be wrong. The multiplier is a label you choose yourself, from memory, about how much you moved this week. Nobody has measured your week. That is the whole error budget of the system in one sentence: the part that is estimated well is a small part of the total, and the part that dominates the total is a guess.

Look at how much of each person's total the basal figure actually covers, because it tells you how much of the number is arithmetic and how much is a guess. For Person A, training moderately, basal is 65 per cent of the total — the other 35 per cent is the multiplier's contribution. For Person B, at a desk with almost no structured activity, basal is 83 per cent, because there is barely any activity component left for the multiplier to get wrong. For Person C, training lightly at a heavier bodyweight, it is 73 per cent. The less active someone's week genuinely is, the more of their number the equation actually knows; the more active they are, the more of it rests on a label they chose from memory about a week nobody measured.

Person B's situation makes the same point from the other side. Her multiplier of 1.2 is not a rounding detail — it is the difference between a plan she can hold and one she quietly breaks by Thursday. If she is genuinely more active than "desk job, little exercise" describes — she walks the dog twice a day, she takes the stairs, she is on her feet more than she credits herself for — her real multiplier sits higher on the table, and her true maintenance is higher than the 1,637 the calculator gave her by an amount that matters at a target this size. She will not find that out from a calculator. She will find it from the scale, over two weeks, eating honestly.

What the number does not tell you

TDEE is a single daily calorie figure. It is not a breakdown of where that energy goes, and it is not a prediction of what happens when you eat less than it.

It does not tell you protein. The number says how much to eat to hold weight. It says nothing about how much of that should be protein, and that is where the difference between losing fat and losing muscle is decided. Set protein separately, after you have the calorie target.
It does not tell you how you will actually lose. A 500-calorie deficit does not come off as clean, even, measurable fat. It comes off as a mix of fat, a little water, a little glycogen, and the exact split shifts week to week. The scale does not separate them for you.
It does not track your real movement. Your maintenance number is fixed in the calculator. Your actual daily expenditure is not. A week where you carry boxes, stand at a market, or skip the lift changes your true number by more than most people's basal rate moves in a decade. The calculator does not know that, and it will not update itself.
It does not tell you when to stop. The estimate was built for your weight today. Every kilogram you lose, the number falls. A target that was correct at 80 kg is too low at 72 kg, and the gap between the two is not small. Recalculate when the scale has moved, not when the calendar says so.

Failure cases: where the estimate misleads and what to do instead

1. "I eat about what the calculator says and I'm not losing anything"

Usually the estimate is fine and the eating is not. "About what the calculator says" is very often somewhat higher in practice — a snack that was not logged, a sauce that was not weighed, a second pour that was not counted, each small enough to feel irrelevant and large enough in total to erase the deficit. One honest week of tracking, weighing food rather than estimating, will tell you which side of the gap you are on. If the tracking shows you are genuinely eating at the target and still not losing, then the estimate was too low and you need to move it up. But do not skip the tracking step to save time; it is the only way to know which of the two explanations is yours.

2. "The number I got is way different from my friend's"

If your inputs are genuinely different — different weight, height, sex, activity level — the numbers should be different, and that is correct. But if your inputs are similar and your numbers are far apart, check the activity row you each picked. That is where most "different number" disputes live. Two people, same body, one picks 1.2 and the other picks 1.725, and the totals differ by more than a third. The calculator did not make a mistake; one of you mislabeled your week.

3. "I hit my calorie target for three weeks and the scale hasn't moved"

Three weeks of honest tracking with no movement is a signal that the estimate was too low by more than the scale's noise can hide. Do not respond by cutting the target further and hoping. Recalculate from your current weight, check your activity level honestly, and if the number is still not moving, accept that your true maintenance is higher than the equation gives and trim the target down by a modest, deliberate amount for the next two weeks — enough to feel like a genuine change, not a repeat of the same guess. The goal is to find the real number, not to punish the estimate for being a guess.

4. "I'm at a maintenance target and I'm still losing weight"

If you are eating at your estimated maintenance and still dropping, either the estimate was slightly too high and you are in a small deficit you didn't intend, or your movement has quietly fallen below what the multiplier assumes. Neither is a failure. If you want to hold weight, eat a little more, or put the extra movement back in — the lift, the walk, the steps you stopped doing after a hard week — and hold the food where it is. If losing is what you want, you found a working number faster than you planned, so keep it and stop second-guessing it. The estimate is a starting point, not a contract, and the scale is the only one allowed to overrule either of you.

How to find your real number

Two weeks, and the method is simple. The point is not to be more disciplined than the equation; it is to let your own body answer a question the equation can only guess at.

Eat at the estimated maintenance figure, consistently, without trying to lose anything. Log every item, weigh food where you can, and do not swap "I'll make it up later" for tracking on the days it is inconvenient.
Weigh yourself every morning, after the toilet, before food, in the same clothes or no clothes — whatever is consistent. The number itself matters less than the trend.
At the end of week one, average the seven morning weights. At the end of week two, average again. Compare the two averages, not any single day.

Steady — within about 0.2 kg between the two weekly averages — means the estimate was close. That is a good outcome; you now know your number and can set a deficit with confidence. Down by roughly half a kilogram means you were eating about 500 calories under your true maintenance the whole time, which is exactly what you wanted to find. Up means the reverse: the estimate was too low and your real number is higher. In every case the scale has given you a number that no equation could, because it came from you.

One refinement: if you are tracking to a deficit rather than at maintenance, divide the weekly average change by 7,700 calories per kilogram of fat to estimate your actual daily deficit, then compare it with what you set out to hit. If Person A's two-week average fell by exactly 0.2 kg, his delivered deficit matches his 220-calorie target — the plan is running as designed. If it fell by less, or by more, the gap between what he set and what actually happened tells him where his real maintenance sits, which is more useful than the estimate ever could be.

Why weekly averages, not daily weights

Daily bodyweight moves by a kilogram or more for reasons that have nothing to do with fat: salt, water, carbohydrate stored as glycogen, what is currently in your digestive tract, and how hard you trained. Reading single mornings is how people conclude that a working plan is failing on day four.

The number changes as you do

Lose weight and your maintenance falls — a smaller body costs less to run and moving it costs less. This is not your metabolism breaking; it is arithmetic. Recalculate every four to six kilograms rather than assuming the number from January still applies in May.

What to actually do with it

Set a deficit that matches a rate of loss you can hold rather than one that looks impressive — the TDEE and calorie target calculator lets you pick the rate and returns the target. Then set protein, because the size of the deficit determines how fast the weight comes off and protein determines how much of it is fat.

One thing worth saying plainly

If your estimate says 2,400 and you have been eating "about 1,800" for months without the scale moving, the likeliest explanation is not a broken metabolism. It is that "about 1,800" is not 1,800. Tracking honestly for a single week, weighing food rather than estimating, resolves this for most people — and it is not something you need to keep doing forever.

The equation this page's calculator uses, and the study behind the tracking-gap point above, are on the references page.

Questions I get asked

Which is more accurate, Mifflin-St Jeor or Harris-Benedict?

Mifflin-St Jeor is the more common modern default and performs at least as well. Neither is precise enough to be treated as a measurement.

Should I eat back the calories from my watch?

Wearable estimates of exercise burn are frequently high. If you eat them all back, a deficit can quietly disappear. Treat them as directional.