TDEE Calculation: Why Three Formulas Give Three Answers

TDEE — total daily energy expenditure — is the number your calorie target depends on, and yet three standard formulas (Mifflin-St Jeor, Harris-Benedict, Katch-McArdle) can produce answers that differ by 200–400 kcal for the same individual. This is not a rounding error — it is a 15–20% uncertainty range that, if you pick the wrong formula, could mean you are eating at maintenance while thinking you are in a deficit. Understanding why the formulas differ, which is most accurate, and when to bypass all of them is foundational knowledge for anyone tracking calories.

The deeper problem is not the formula — it is the activity multiplier. Every TDEE calculator asks you to self-classify as “sedentary,” “lightly active,” or “moderately active” and multiplies your BMR by a coefficient (1.2, 1.375, 1.55, etc.). These multipliers introduce far more error than the BMR formula itself. Per Daly & Burke 2004 (International Journal of Sport Nutrition and Exercise Metabolism), self-reported activity levels systematically overestimate actual activity, often by 1–2 activity tiers — adding 200–400 kcal/day to the calculated TDEE without any basis in reality.¹

CalEye calculates your actual TDEE from your logged calorie intake and real-world weight trend — bypassing formula uncertainty entirely with empirical data.

Mifflin-St Jeor: The Current Clinical Standard

The Mifflin-St Jeor equation was published in 1990 and is currently the formula recommended by the Academy of Nutrition and Dietetics for estimating resting metabolic rate (RMR) in non-obese adults.² It uses height, weight, age, and sex:

For men: BMR = (10 × weight in kg) + (6.25 × height in cm) − (5 × age in years) + 5

For women: BMR = (10 × weight in kg) + (6.25 × height in cm) − (5 × age in years) − 161

For a 75 kg, 175 cm, 35-year-old man: BMR = (10 × 75) + (6.25 × 175) − (5 × 35) + 5 = 750 + 1,093.75 − 175 + 5 = 1,674 kcal/day.

When validated against doubly labelled water studies — the gold standard for measuring actual energy expenditure — Mifflin-St Jeor demonstrates a mean prediction error of approximately ±10% across diverse adult populations.³ That 10% figure is important: for an individual with a true BMR of 1,700 kcal/day, the formula might output anything from 1,530 to 1,870 kcal/day and still be within its stated accuracy range. This is not a flaw unique to Mifflin-St Jeor — it reflects the fundamental biological reality that two people with identical anthropometric inputs can have meaningfully different resting metabolic rates due to differences in organ mass, mitochondrial efficiency, thyroid function, and body composition within the same body weight category.

The formula’s significant limitation is that it uses total body weight rather than lean body mass. Two people at the same weight, height, age, and sex can differ by 10–15 kg in lean mass — and since muscle is metabolically more active than fat, the high-lean-mass individual will have a higher actual BMR than Mifflin-St Jeor predicts. This is where Katch-McArdle becomes relevant.

Harris-Benedict Revised: Older but Still Common

The Harris-Benedict equation was originally published in 1919, making it one of the oldest BMR formulas still in clinical circulation. The original version was revised by Roza and Shizgal in 1984 using a more modern validation dataset, and the revised version is what most clinical references mean when they cite “Harris-Benedict.”⁴

Revised Harris-Benedict for men: BMR = (13.397 × weight in kg) + (4.799 × height in cm) − (5.677 × age in years) + 88.362

For the same 75 kg, 175 cm, 35-year-old man: BMR = (13.397 × 75) + (4.799 × 175) − (5.677 × 35) + 88.362 = 1,004.8 + 839.8 − 198.7 + 88.4 = 1,734 kcal/day.

That is 60 kcal/day higher than Mifflin-St Jeor for identical inputs — a small but consistent upward bias that compounds to 420 kcal per week, or the rough equivalent of one extra half-pound of fat loss per month at a targeted deficit. If you are using Harris-Benedict to set a calorie target and your actual BMR is closer to the Mifflin-St Jeor estimate, you are eating 60 kcal/day more than you intend.

Multiple comparative studies have found that the revised Harris-Benedict equation overestimates BMR by approximately 5% versus doubly labelled water measurements in most adult populations — a consistent directional bias, not random error.³ This means it systematically calculates a higher BMR than you actually have. The formula remains widely used in clinical settings due to decades of historical familiarity and its presence in medical software, not because it is more accurate than Mifflin-St Jeor. If you have access to both, default to Mifflin-St Jeor.

Katch-McArdle: The Lean-Mass Formula

Katch-McArdle takes a different approach: it calculates BMR from fat-free mass (FFM) rather than total body weight, which is theoretically more accurate because metabolically active tissue — muscle, organs — drives resting energy expenditure, not adipose tissue.

Katch-McArdle: BMR = 370 + (21.6 × fat-free mass in kg)

For a 75 kg person with 15% body fat, FFM = 75 × 0.85 = 63.75 kg. BMR = 370 + (21.6 × 63.75) = 370 + 1,377 = 1,747 kcal/day — meaningfully higher than both Mifflin-St Jeor (1,674) and Harris-Benedict (1,734) for the same total weight.

Now change the body fat percentage to 25% (same total weight, different composition): FFM = 75 × 0.75 = 56.25 kg. BMR = 370 + (21.6 × 56.25) = 370 + 1,215 = 1,585 kcal/day — 89 kcal lower than Mifflin-St Jeor. This divergence is the formula doing its job: correctly accounting for the fact that a leaner body at the same weight burns more calories at rest.

The limitation is obvious: Katch-McArdle requires an accurate body fat percentage measurement as input. DEXA scan is the most accurate consumer-accessible method (error rate ±1.5–2.5% body fat).⁵ Bioelectrical impedance (the scale at your gym) has an error rate of ±3–8% body fat depending on hydration state, time of day, and device quality. Skin-fold caliper measurement by a trained practitioner is ±3–5% body fat. Substituting an inaccurate body fat measurement into Katch-McArdle produces less reliable results than simply using Mifflin-St Jeor. Katch-McArdle earns its place for athletes who know their lean mass via DEXA; it is not useful for the general population without that input.

The Activity Multiplier: The Real Source of Formula Divergence

All three formulas produce BMR estimates within 5–10% of each other for the same person and the same inputs. The truly large divergences in final TDEE come not from the formula choice but from the activity multiplier applied afterward.

The standard multipliers:

• Sedentary (desk job, no deliberate exercise): BMR × 1.2
• Lightly active (1–3 workouts per week): BMR × 1.375
• Moderately active (3–5 workouts per week): BMR × 1.55
• Very active (hard training 6–7 days per week): BMR × 1.725
• Extra active (physical job plus training): BMR × 1.9

Applying 1.55 versus 1.375 on an 1,800 kcal BMR produces TDEE estimates of 2,790 versus 2,475 kcal — a 315 kcal difference from one tier of self-classification error. This dwarfs the 60–90 kcal difference between formulas. And the direction of error is systematic: most people self-classify one tier higher than their actual activity warrants. A person who goes to the gym three times per week for 45 minutes of mixed cardio and weights often self-identifies as “moderately active” (1.55 multiplier) when their non-exercise movement (steps, standing time, incidental activity) is sedentary enough to put them closer to “lightly active” (1.375).¹

The consequence is systematic TDEE overestimation. Someone who calculates TDEE as 2,790 kcal/day and sets a 500 kcal deficit at 2,290 kcal/day may in reality have a true TDEE of 2,475 kcal/day — meaning their “500 kcal deficit” is actually 185 kcal. At that deficit, weight loss would be approximately 0.17 kg per week rather than the expected 0.45 kg per week. After 4–6 weeks of unexpectedly slow progress, many people conclude that “their metabolism is broken” rather than recognizing that the activity multiplier was the source of error.

Empirical TDEE: The Superior Method

Rather than using a formula, the most accurate TDEE estimate comes from 3 weeks of logged food intake combined with weekly average body weight. This empirical approach eliminates all formula uncertainty because it measures what you are actually expending — indirectly, via the relationship between intake and weight change.

The logic: if body weight is completely stable over 3 weeks while you average 2,200 kcal/day of logged intake, your TDEE is approximately 2,200 kcal/day. The body doesn’t manufacture or destroy energy; if weight is not changing, intake equals expenditure by definition.

If weight is falling at 0.3 kg per week on average intake of 1,900 kcal/day, your implied deficit is approximately 300 kcal/day (0.3 kg × ~1,000 kcal per kg of body mass, divided by 7 days), placing your TDEE at approximately 2,200 kcal/day. The calculation is the same regardless of whether you are “sedentary” or “moderately active” — real-world intake and weight trend data capture your actual total expenditure without requiring any self-classification.

This approach requires two things: accurate logging and patience. Logging accuracy matters — underreporting food intake by 20% (a well-documented pattern) would make a 2,200 kcal TDEE appear to be a 1,760 kcal TDEE, leading to a too-aggressive initial deficit.⁶ This is why photo-based logging with USDA-sourced references is more reliable than memory-based estimates for the empirical calibration period. Three weeks of consistent, accurate logging produces an empirical TDEE estimate that is almost always more accurate than any formula — because it is specific to your body, your current activity level, your basal metabolic rate, and your measurement habits.

When Formula Estimates Are Useful and When to Abandon Them

Use formula-based TDEE estimates as a starting point only. They provide a reasonable initial calorie target before you have any real-world data — something you need on day 1, when three weeks of empirical data don’t yet exist. Mifflin-St Jeor at the “lightly active” multiplier (1.375) is the most defensible starting point for most people, because it uses the most validated formula and selects a conservative activity tier that most people will meet or exceed rather than fall short of.

Abandon formula estimates as soon as you have 2–3 weeks of accurate intake logging and daily weigh-in data. At that point, the empirical estimate is simply more accurate, and there is no reason to continue using an approximation when you have direct measurement. Set your calorie target based on the empirical TDEE minus your intended deficit (typically 300–500 kcal for sustainable fat loss, 200–300 kcal for preservation of lean mass during a slow cut).

Revisit the empirical calculation every 4–6 weeks. As body weight falls, TDEE decreases — a phenomenon called metabolic adaptation, where weight loss itself reduces resting metabolic rate, both because a lighter body requires less energy to move and because some degree of adaptive thermogenesis (a downward adjustment in non-exercise activity thermogenesis) typically occurs with sustained calorie restriction.⁷ Recalculating every 4–6 weeks from fresh weight-trend and intake data keeps the calorie target calibrated to your current physiology rather than the physiology you had at the start.

References

1. Daly RM, Burke LM. “Self-Reported Physical Activity as a Predictor of Actual Activity Levels.” International Journal of Sport Nutrition and Exercise Metabolism 14, no. 5 (2004): 578–594.

2. Mifflin 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 51, no. 2 (1990): 241–247.

3. Frankenfield 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 105, no. 5 (2005): 775–789.

4. Roza AM, Shizgal HM. “The Harris Benedict Equation Reevaluated: Resting Energy Requirements and the Body Cell Mass.” American Journal of Clinical Nutrition 40, no. 1 (1984): 168–182.

5. Shepherd JA, Ng BK, Sommer MJ, Heymsfield SB. “Body Composition by DXA.” Bone 104 (2017): 101–105.

6. Dhurandhar NV, Schoeller D, Brown AW, et al. “Energy Balance Measurement: When Something Is Not Better Than Nothing.” International Journal of Obesity 39, no. 7 (2015): 1109–1113.

7. Rosenbaum M, Leibel RL. “Adaptive Thermogenesis in Humans.” International Journal of Obesity 34, Supplement 1 (2010): S47–S55.