Adaptive TDEE: why calorie calculators get it wrong — and how to find your real energy needs

Bilberry Guide · updated: 22 Jul 2026 · approx. 9 min read

Two people with the same height, weight and age type their stats into a calorie calculator and get an identical result — say 2,300 kcal. After two months of eating "by the number", one is losing weight on schedule and the other hasn't moved. Neither of them is cheating. The calculator simply doesn't know their real metabolism — because it can't.

What TDEE actually is

TDEE (Total Daily Energy Expenditure) is the total number of calories your body burns in a day. It has four components:

Calculators estimate BMR from a formula (usually Mifflin-St Jeor or Harris-Benedict), then multiply it by an activity factor (PAL) that you declare yourself. And that's where the trouble starts.

Three reasons the formula misses

1. The BMR formula is a population average. The equations were built by fitting measurements from large groups of people. For the average person in that group they work reasonably well, but individually, deviations of 10–20% are completely normal — and 15% of 1,800 kcal is already nearly 300 kcal a day. In either direction.

2. NEAT varies between people more than anything else. Differences in spontaneous activity between people with similar lifestyles can reach several hundred kcal a day. Worse, NEAT changes on its own: in a calorie deficit your body quietly dials it down — you fidget less, sit down more often, walk more slowly. No formula sees any of that.

3. The activity factor is a guess. "Moderately active" means three workouts a week and a standing job to one person — and walking the dog to another. Choosing between 1.4 and 1.6 shifts the result by several hundred kcal, and you pick it by eye.

Adaptive TDEE: measure your metabolism from your own data

There is a way around all of these problems at once — don't estimate TDEE from a formula, measure it from what is actually happening. You need two things you already have on hand: a food log and a scale.

The logic is simple: if you know your average calorie intake and how your weight has changed, you can run the energy balance backwards:

TDEE ≈ average daily intake − (weight change [kg] × ~7,700 kcal) / number of days

An example: you eat 2,000 kcal a day on average, and over 3 weeks (21 days) you've lost 0.7 kg. The deficit was 0.7 × 7,700 ≈ 5,400 kcal, i.e. ~260 kcal a day. Your real TDEE is about 2,260 kcal — no matter what any formula claims.

Two details decide whether this calculation works:

A weight trend instead of single weigh-ins. Day to day, your weight jumps by ±1–2 kg from water, salt, glycogen and gut contents. A single weigh-in tells you nothing; what counts is the smoothed trend line drawn from daily measurements (a moving or exponential average). Only its slope is real tissue change.

A window of at least 2–3 weeks. Shorter periods are dominated by water noise. The longer the window (and the more honest the log), the more accurate the result — and logging errors (underestimated portions) partly cancel out of the math anyway, because they're systematic.

What throws the math off: noise worth knowing about

Adaptive TDEE is only as good as the data you feed it. A few things can temporarily "move" the scale without any change in body fat — and if you don't know about them, you'll draw the wrong conclusions:

Salt and carbohydrates. A big, salty dinner or a higher-carb day can mean +1–2 kg the next morning — that's water bound to glycogen and sodium, not fat. It's gone in 2–3 days.

Creatine. Starting supplementation adds 1–2 kg of water inside your muscles. It's a one-off shift — the trend returns to normal, but a measurement window that includes the start of creatine will understate your computed TDEE.

The menstrual cycle. Water retention across the phases of the cycle can reach 1–3 kg. The fairest comparison is the same week of the cycle, month to month — or simply a measurement window one full cycle long.

Alcohol and stress. Alcohol dehydrates first, then promotes water retention; elevated cortisol does too. A trip, a rough week at work — the scale "stalls" even though fat is dropping. The trend straightens it out, as long as you give it time.

The most common mistakes in adaptive tracking

A window that's too short. A week of data is a lottery — one salty dinner at the end of the window is all it takes. Minimum 2–3 weeks; 4 if your routine is irregular.

Skipping weekends. The most common mistake in practice: a log from Monday to Friday, the weekend "doesn't count". It counts — that's often exactly where the deficit disappears. The math has to cover full weeks, otherwise your average intake is fiction.

Changing how you log mid-window. If halfway through the window you start weighing food instead of eyeballing it, your "intake" jumps without any change in your diet. Start a fresh measurement window after every such change.

Weighing at different times. In the morning, after the bathroom, before eating — always the same way. An evening weigh-in can run 1–2 kg above the morning one.

Can you raise your TDEE?

Yes — though more slowly than the headlines promise. Three real levers, in order of effectiveness:

NEAT is the biggest lever. An hour of walking is 200–350 kcal; a standing desk, stairs instead of the elevator, phone calls taken on foot — that's hundreds of kcal a day, available right now and with no training fatigue.

Muscle mass works, but modestly. A kilogram of muscle burns roughly 10–15 kcal a day at rest — building muscle has plenty of benefits, but "revving up your metabolism" by 500 kcal through lifting alone is a myth. The training itself and the recovery after it do cost energy, though.

After a cut: raise calories gradually. Coming back to maintenance eating after a long deficit is best spread over a few weeks (+100–150 kcal per week) while watching the trend — the body partly rebuilds NEAT and TDEE genuinely rises.

How to do this in practice

By hand: log your calories every day, weigh yourself every morning under the same conditions, and every 2–3 weeks compute your average intake and the slope of your weight trend, then plug them into the formula above. A spreadsheet is enough — it only takes consistency and a little arithmetic.

Automatically: Bilberry runs exactly this calculation for you. You log meals (including by photo — the AI recognizes the ingredients and calories), enter your weight, and the app keeps a smoothed weight trend and adjusts your TDEE and calorie target every week based on your actual results. The calculator formula serves only as a starting point for the first two weeks — after that, the only thing that matters is what your data shows.

Frequently asked questions

How many calories a day to lose weight? A safe pace is usually losing 0.5–1% of body weight per week, which corresponds to a deficit of roughly 300–600 kcal below TDEE. The key is knowing your true TDEE — calculate it adaptively; how to choose the deficit itself is covered in a separate article.

Do I have to weigh myself every day? You don't, but daily weigh-ins give the smoothest trend and detect changes fastest. Weighing once a week works too — it just takes more weeks to get a confident answer.

What if I'm not losing weight despite a deficit? Most often the deficit only exists on paper: actual intake is higher than logged, or TDEE is lower than the calculator said. The adaptive calculation settles it in 2–3 weeks — it shows your actual TDEE, and then it's immediately clear where to set the target.

Is 7,700 kcal per kilogram a reliable number? It's an approximation for fat tissue with a small lean-mass component; the true value depends on the composition of the weight you lose. For weekly target adjustments it's accurate enough.