ENERGY ANALYSIS

Adaptive TDEE

Estimate your Total Daily Energy Expenditure (TDEE) from the relationship between logged energy intake and observed body-weight change.

Open Adaptive TDEE in Fuel Lens ↗

What TDEE means

TDEE is the estimated number of kilocalories your body uses per day. It includes resting metabolism, normal daily movement, digestion, and exercise when the selected model or dashboard configuration includes it.

The central energy-balance idea is:

Energy stored = Energy eaten − Energy expended

Rearranged for expenditure:

TDEE ≈ average intake − energy equivalent of observed weight change

A stable weight at 2,800 kcal/day suggests a TDEE near 2,800 kcal/day. Losing weight while eating 2,800 kcal/day suggests that TDEE is higher than 2,800; gaining weight suggests that it is lower.

Adaptive TDEE is an estimate, not a laboratory measurement. Water, glycogen, food in the digestive system, sodium, menstrual-cycle changes, illness, and logging errors can temporarily move scale weight without representing an equivalent change in stored tissue.

Which model is used?

Fuel Lens supports two models. The default is First Principles TDEE.

Model	How it works	Best use
First Principles TDEE	Estimates expenditure directly from intake and the smoothed weight trend in overlapping rolling windows	The primary adaptive estimate when enough real observations exist
Statistical TDEE	Starts with a BMR-based prior and blends it toward the intake/weight estimate according to data coverage	A useful comparison or fallback when data is sparse

Legacy names dynamic, rolling-window, and hybrid are normalized to First Principles TDEE.

First Principles TDEE

1. Build the daily timeline

The calculator creates one daily timeline covering the first through last date found in energy, exercise, or weight data.

Energy labels recognized by default: Energy, Calories, and Calories (kcal).
If several recognized labels exist on one day, the first matching label in the configured label order is used; values are not added together.
Exercise calories are summed per day and are used for diagnostics and optional baseline calculations.
Multiple weight entries on the same day are averaged.
Missing weights between two observations are linearly interpolated.
The interpolated series is smoothed with a 7-day moving average to reduce daily scale noise.

2. Estimate each rolling window

The default rolling window is 21 days, moving forward in 7-day steps. At least 14 timeline days are needed before First Principles can run.

For each window, the calculator requires:

at least max(2 × minimum calorie days, 50% of the window) intake days;
at least max(minimum weight entries + 2, 20% of the window) observed weight days; and
the same minimum number of smoothed weight points.

The default minimums are 4 calorie days and 2 weight entries. A window that does not meet them is rejected.

The smoothed weights are fitted with a least-squares linear regression. If wᵢ is smoothed weight on timeline day i, the slope is:

weight slope = Σ(i − mean(i))(wᵢ − mean(w)) / Σ(i − mean(i))²

The slope is measured in kg/day. A negative slope means weight is falling.

3. Convert weight change into energy

The model uses an effective energy density rather than assuming that every kilogram is pure fat.

Constants used by the implementation:

fat energy density = 9,300 kcal/kg

lean-tissue energy density = 1,100 kcal/kg

If body-fat percentage is available, the estimated fat share is:

fat share = clamp(0.55 + 0.40 × body-fat fraction, 0.55, 0.85)

Without body-fat data, the default fat share is 0.66. The baseline energy density is:

baseline density = fat share × 9,300 + (1 − fat share) × 1,100

Rapid weight changes receive a penalty because they are more likely to include water and glycogen shifts:

weekly change = |weight slope| × 7

rapid-shift factor = clamp(1 − 0.08 × weekly change, 0.75, 1.00)

effective density = round(baseline density × rapid-shift factor)

4. Calculate each window's TDEE

For a window with average daily intake I, weight slope S in kg/day, and effective density D in kcal/kg:

window TDEE = I − D × S

This has the expected direction:

If weight is stable, S ≈ 0, so TDEE is close to intake.
If weight is falling, S < 0, so the subtraction adds calories to intake.
If weight is rising, S > 0, so the estimate is below intake.

Example using the simple 7,700 kcal/kg approximation:

intake = 2,400 kcal/day

weight change = −1.0 kg over 21 days = −0.0476 kg/day

TDEE ≈ 2,400 − 7,700 × (−0.0476) ≈ 2,767 kcal/day

The production calculation may use a different effective density based on body composition and rate of change.

5. Combine windows robustly

All valid window estimates are collected. If there are at least six estimates, the lowest and highest 15% are trimmed. The final First Principles TDEE is the median of the remaining estimates:

final TDEE = median(robust window estimates)

The trend slope shown in diagnostics is the regression trend across the window estimates, expressed in kcal per window. The estimate is rounded to whole kcal/day.

Statistical TDEE

Statistical TDEE is a BMR-based prior that adapts toward observed energy balance.

1. Calculate the BMR prior

The calculator chooses the first available option in this order.

Katch–McArdle

Used when weight and body-fat percentage are available:

lean mass = weight × (1 − body-fat percentage / 100)

BMR = 370 + 21.6 × lean mass

Weight is in kg and lean mass is in kg.

Mifflin–St Jeor

Used when weight, height, age, and sex are available:

BMR = 10 × weight + 6.25 × height − 5 × age + s

Height is in cm. The sex constant is:

Sex value	`s`
Male	`+5`
Female	`−161`

The implementation treats a value beginning with m as male and other available values as female for this formula.

Configured fallback BMR

If neither formula can be calculated, the configured fallbackBmr is used. If no valid BMR source exists, Statistical TDEE returns no result.

2. Apply activity assumptions

The prior TDEE is:

prior TDEE = BMR × activity factor

Supported activity levels map to these factors:

Activity level	Factor
No activity	`1.00`
Sedentary	`1.20`
Lightly active	`1.375`
Moderately active	`1.50`
Very active	`1.90`
Custom	`1 + custom activity kcal / BMR`

If no activity level or custom factor is supplied, the default factor is 1.20.

Exercise calories are not added by default. When countExerciseAdditively is enabled, average exercise calories are added to the fallback prior. The UI also reports exercise separately so that baseline expenditure can be shown without double-counting it.

3. Calculate the observed TDEE

When there is enough intake and weight data, Statistical TDEE uses the same smoothed regression and effective-density calculation as First Principles:

observed TDEE = average intake − effective density × weight slope

Evidence is based on calorie and weight coverage:

calorie coverage = clamp(calorie days / timeline days, 0, 1)

weight coverage = clamp(weight points / (timeline days / 3), 0, 1)

evidence score = clamp(0.65 × calorie coverage + 0.35 × weight coverage, 0, 1)

The blend weight is:

blend weight = clamp(0.15 + 0.70 × evidence score, 0.15, 0.85)

4. Blend and converge

Starting from the prior, the model moves toward the observed estimate:

next TDEE = current TDEE + (observed TDEE − current TDEE) × blend weight

The result is rounded after each iteration. Iteration stops when the change is at most 1 kcal/day or after 16 iterations:

statistical TDEE = converged current TDEE

With sparse data, the model skips the observed estimate and returns the activity-adjusted prior. Diagnostics identify this as a fallback.

Exercise and baseline values

The primary TDEE estimate is based on gross logged intake and observed weight change. Exercise is not subtracted from the First Principles formula.

The UI separately reports:

average net intake = average intake − average exercise calories

non-exercise TDEE = TDEE − average exercise calories

These are interpretation values. Do not subtract exercise from intake and then subtract it again from TDEE; that would count the same activity twice.

Defaults, limits, and fallback behavior

Setting	Default
Model	First Principles
Minimum timeline	14 days
Rolling window	21 days
Window step	7 days
Minimum calorie days	4
Minimum weight entries	2
Accepted TDEE range	500–10,000 kcal/day
Statistical activity factor	1.20
Add exercise to Statistical fallback	No

The calculator falls back to Statistical TDEE when First Principles has fewer than 14 days, has no valid rolling windows, or produces a value outside the accepted range. Statistical TDEE can also fall back to its BMR/activity prior when intake or weight evidence is insufficient, invalid, or outside the accepted range.

Understanding the result

Total Energy Expenditure: the estimated daily energy use.
Baseline Metabolic Rate: the BMR-based or non-exercise baseline shown by the selected view; it is not the same thing as TDEE.
Average Intake: average daily energy from the selected recognized energy label.
Average Net Intake: average intake minus logged exercise calories.
Weight Change: observed change and its regression-derived rate.
Activity Factor: the effective TDEE/BMR ratio in the statistical model.
Confidence: a qualitative signal based on calorie coverage, weight coverage, and the number of valid windows.
Current Target: the active dashboard calorie target. It is a planning value, not necessarily the adaptive estimate.
Calorie Gap:

calorie gap = average intake − TDEE

A positive gap indicates intake above the estimate; a negative gap indicates intake below it.

How to get a better estimate

Log intake consistently, including weekends and meals away from home.
Record weight under similar conditions and use several observations per week when possible.
Keep the selected date range long enough to include multiple rolling windows.
Avoid interpreting a single day's result; use the estimate and its trend over time.
Treat sudden changes after high-sodium meals, travel, illness, or unusual training as potentially transient.
Check the displayed confidence, calorie coverage, weight coverage, and whether a fallback was used.

Adaptive TDEE becomes more useful as consistent intake and weight observations accumulate. It should support decisions about calorie targets, not replace medical or clinical assessment.

Scientific background

The intake-and-weight-change approach follows energy-balance accounting, but human weight change is dynamic: energy expenditure and body composition can adapt over time. Hall and colleagues describe why static calorie-to-weight rules do not predict the time course reliably. A Fuel Lens estimate is a practical estimate from logged data, not a laboratory measurement or a reproduction of their full physiological model. Self-reported intake and short-term body-weight changes also carry measurement error.

Hall KD et al. (2011). Quantification of the effect of energy imbalance on bodyweight. The Lancet, 378(9793), 826–837.
Racette SB et al. (2007). Validation study of energy expenditure and intake during calorie restriction using doubly labeled water and changes in body composition. The American Journal of Clinical Nutrition, 85(1), 73–79.

Continue with these connected Fuel Lens guides.