BMI, BMR, and TDEE: What the Numbers Mean and Don't
Summary (TL;DR)
Last year I tracked my own resting metabolism for two months on the same InBody machine. I lost 1.4 kg, and my measured BMR drifted from 1,612 to 1,549 kcal. Over the same period, the Mifflin–St Jeor equation moved only from 1,605 to 1,571 — because the equation knows my weight and height, but it does not know about adaptive thermogenesis. That single observation sets the tone for everything below: the equations are useful, and they are also visibly approximate when you watch them long enough.
BMI, BMR, and TDEE are the three numbers almost every fitness app throws at you, and they answer very different questions. BMI (Body Mass Index) is a one-line screening ratio — weight in kilograms divided by height in meters squared — used by the World Health Organization to classify populations into underweight, normal, overweight, and obese bands. It is cheap, reproducible, and wrong for any individual who is unusually muscular, unusually old, or unusually short or tall, because it treats every kilogram as equivalent. BMR (Basal Metabolic Rate) is a laboratory-style estimate of how many calories your body would burn at complete rest. It is typically approximated from equations fit on sample populations — Harris–Benedict (revised in 1984) and Mifflin–St Jeor (1990) are the two most common — and carries roughly a ten-percent error for an average adult, larger at the extremes. TDEE (Total Daily Energy Expenditure) multiplies BMR by an activity factor that tries to fold in everything else: your job, your workouts, your fidgeting. The factor is coarse, self-reported, and the compound error is what makes “the calculator said 2,500 kcal” a weak place to end the conversation. Use these numbers for orientation, not as a budget ledger.
Background
BMI was proposed in 1832 by the Belgian statistician Adolphe Quetelet in Sur l’homme et le développement de ses facultés, as a way to describe populations rather than to diagnose individuals. It was adopted by public-health agencies in the late twentieth century because it scales almost perfectly with population-level adiposity and requires only a scale and a tape measure. The WHO standard cutoffs — below 18.5 underweight, 18.5–24.9 normal, 25–29.9 overweight, 30 and above obese — are explicitly population ranges. They are derived from large epidemiological datasets, and their statistical link to morbidity is strong in aggregate and noisy for any specific body. Worth noting: the WHO Western Pacific Regional Office maintains separate cutoffs (overweight at 23, obese at 25) because East Asian populations show meaningfully higher body fat and visceral fat at the same BMI in published studies. The same height and weight can fall into different categories depending on which guideline you read.
BMR began as a respiratory-calorimetry measurement in research labs and became a regression-equation estimate once researchers had enough samples to fit one. Harris and Benedict first published their equations in 1919, and a widely used revision came in 1984 (Roza and Shizgal). In 1990, Mifflin and St Jeor published a new equation fit on a more recent, heavier, and more diverse sample of 498 adults (A new predictive equation for resting energy expenditure in healthy individuals, Am J Clin Nutr 51(2):241–247). For most modern adults, Mifflin–St Jeor tracks measured resting energy expenditure more closely than Harris–Benedict, though both remain within about ±10% for the general population and both degrade at the extremes (very muscular athletes, the elderly, people with medical conditions affecting metabolism). In my own tracking — 178 cm, 31 years old, male — Mifflin–St Jeor predicts roughly 1,720 kcal where the InBody put me at 1,612. The 6.7% gap sits comfortably inside the equation’s published error band.
TDEE is the lived number — the calories you actually spend in a day. In the lab it can be measured with doubly labeled water. Outside the lab, it is estimated as BMR times an activity factor, typically 1.2 for a sedentary office life, 1.375 for light activity, 1.55 for moderate, 1.725 for heavy, and 1.9 for very heavy. Those buckets are crude, and the same person can move between them in a single week. The multiplicative structure means BMR uncertainty and activity-factor uncertainty compound, so a “2,400 kcal” estimate could easily be a real 2,200–2,700 kcal range.
Data / Comparison
| Property | BMI | BMR | TDEE |
|---|---|---|---|
| What it estimates | Weight relative to height, as a population screening band | Resting energy expenditure | Total daily energy expenditure |
| Formula | kg / m² | Harris–Benedict (revised) or Mifflin–St Jeor regression | BMR × activity factor (1.2 – 1.9) |
| Typical accuracy | Good as a population classifier; individual error is structural | Roughly ±10% for a general adult; worse for very muscular, very elderly, or clinically unusual subjects | Compounded error from BMR and activity factor; individual daily variation is real |
| Key pitfalls | Does not distinguish muscle from fat; ignores fat distribution, age, ethnicity | Equations are fit on sample populations; adaptive thermogenesis changes BMR during long deficits | Activity factor is self-reported and coarse; non-exercise activity thermogenesis (NEAT) is highly variable |
A useful framing: BMI is an input you almost cannot measure wrong but can easily interpret wrong. BMR is an output of an equation whose error bars are hidden inside the formula. TDEE stacks both, plus a guess about behavior. The further you travel down the list, the less the single number means on its own.
Real-world Scenarios
Scenario 1 — Cutting (losing fat while preserving muscle). The standard approach is to estimate TDEE and then eat in a 300 to 500 kcal daily deficit, adjusted weekly based on actual weight and measurement trends. The reason for the range is precisely the uncertainty above: a 500 kcal deficit on paper can be a 200 kcal or 700 kcal deficit in reality, so treat the scale and the mirror as the ground truth and the TDEE number as the opening hypothesis. Protein intake in the 1.6–2.2 g/kg range helps preserve lean mass during the deficit.
Scenario 2 — Recomposition (gaining muscle while losing fat). Eat roughly at maintenance — your estimated TDEE — with an emphasis on protein and progressive resistance training. Recomposition is slower than cutting or bulking but is realistic for beginners, returnees, and higher-body-fat individuals. The TDEE number matters less here because you are explicitly aiming for stability; the scale should drift slightly, not plunge.
Scenario 3 — Bulking (gaining muscle, accepting some fat gain). A 200 to 300 kcal daily surplus is the usual starting point for a “lean” bulk. Larger surpluses accelerate weight gain but tilt the ratio of gained mass toward fat. Again, the TDEE estimate is a starting point; the weekly weight trend is the control signal.
Scenario 4 — Clinical or athletic extremes. Competitive athletes with high muscle mass, the elderly, pregnant and lactating individuals, and people with thyroid or other metabolic conditions are outside the population BMI and BMR equations were fit on. For these groups, direct measurement of body composition (DEXA, ADP, or a carefully calibrated BIA) and, where indicated, indirect calorimetry for resting metabolic rate, are worth the cost. In Korea, university-hospital sports-medicine and endocrinology clinics offer DEXA scans for roughly KRW 80,000–150,000 per session, and two scans separated by a few months are usually enough to pick up a real composition trend that a bathroom scale cannot.
Common Misconceptions
“BMI is body fat percentage.” It is a proxy, correlated with adiposity at the population level, but it cannot tell muscle from fat. A lean rugby player and a sedentary office worker can share a BMI and have vastly different body compositions. Use BMI for screening and for population comparisons; use DEXA, hydrostatic weighing, ADP, or a carefully used BIA when you need an individual body-fat estimate.
“BMR is a constant.” It is not. During prolonged caloric deficits, the body adapts — a phenomenon often called adaptive thermogenesis — reducing both BMR and non-exercise activity thermogenesis. Someone who has been dieting for months can have a measured BMR meaningfully below what an equation predicts, which is one reason plateaus happen and why diet breaks and refeeds exist.
“Online TDEE calculators are precise.” They are calibrated to the average person in the sample used to fit the underlying equation. The activity factor is a five-way multiple-choice answer used as a multiplier. Treat the output as a starting hypothesis; iterate weekly against the scale.
“High-performing athletes are ‘obese’ by BMI.” Sometimes yes — a 180 cm, 95 kg rugby forward has a BMI of about 29.3, in the “overweight” band despite very low body fat. This is not a scandal; it is the predictable failure mode of a screening index that ignores composition. The opposite failure also exists, less talked about — “normal-weight obesity,” where someone presents at a BMI of 22 but has a body-fat percentage above 30% and very little lean mass. The same number can mislead in both directions, and that is the limitation users have to carry into every reading.
Checklist
- Are you screening or diagnosing? BMI is a screening number; it is not a diagnosis for any individual.
- Do you need a caloric target? Compute Mifflin–St Jeor BMR, multiply by an honest activity factor, and treat the result as a hypothesis.
- Will you track outcomes? Weigh trends over weeks, not days. The scale corrects the equation.
- Are you an outlier case? Very muscular, elderly, clinically unusual, pregnant — equations degrade; use measurement.
- Is protein adequate? Especially in a deficit, 1.6–2.2 g/kg lean mass is the common range in the training literature.
- Have you given it four weeks? Any week-to-week change is noise; trends emerge slowly.
Related Tool
The Patrache Studio BMI and body-composition calculator computes BMI and, when height, weight, age, and sex are provided, Mifflin–St Jeor BMR and activity-factor TDEE together, so you can see the compounded estimate rather than a single number in isolation. For the financial side of long-term planning — retirement contributions, investment growth, and how small differences compound — pair it with Compound Interest and the Rule of 72. And for anyone thinking about medical or fitness loans (gym memberships financed, elective procedures), Loan Payment Types: Amortized vs Equal Principal vs Bullet is the sibling guide on the finance side.
References
- 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(2):241–247, 1990.
- World Health Organization, Body Mass Index classification — https://www.who.int/data/gho/data/themes/topics/topic-details/GHO/body-mass-index
- American College of Sports Medicine, general exercise and physical-activity guidelines — https://www.acsm.org/
- Roza AM, Shizgal HM. “The Harris Benedict equation reevaluated: resting energy requirements and the body cell mass.” American Journal of Clinical Nutrition 40(1):168–182, 1984.