TDEE Formula Comparison: Which One Should You Use?

Which TDEE formula should you use?

Use the comparison first, then let logged data decide. For adults 19 and over, the calculator's TDEE comes from the 2023 DRI EER, which predicts total daily burn directly from doubly labeled water research; the BMR it shows is Mifflin-St Jeor. Enter your body fat and both numbers come from Katch-McArdle instead; enter body fat and pick a sport under Sport / Discipline, and they come from Cunningham. Among BMR equations, Mifflin-St Jeor is the best-tested choice when you do not know body fat, and Katch-McArdle or Cunningham is better when lean mass is measured well. The FitnessVolt calculator compares all 8 formula paths, plus strength-aware athlete logic, so you can see the uncertainty range before trusting one number.

• No body-fat data: start with the default (DRI EER) or Mifflin-St Jeor and a realistic activity level.
• Measured body fat or serious lifting: compare Katch-McArdle, Cunningham, and the strength-adjusted estimate.
• Wide formula spread: use the first result as a hypothesis, then tighten it with adaptive calibration; your target can start to move after the second weekly check-in, around day 14.

Why Multiple Formulas Exist

There is no single universally accurate formula for predicting resting metabolic rate, because metabolism is not a simple function of height, weight, age, and sex. Different populations - athletes, older adults, individuals with high body fat, those in food-scarce regions - have metabolic characteristics that are not fully captured by equations derived from specific study cohorts. As a result, researchers have developed multiple formulas over the past century, each optimized for different populations and validated against different datasets.

The practical consequence for you is this: the formula that best fits a 25-year-old male powerlifter (Cunningham or Katch-McArdle) may systematically overestimate BMR for a 55-year-old sedentary woman (for whom WHO/Schofield or Mifflin is better calibrated). Using the wrong formula introduces a predictable bias that undermines calorie tracking for months before you realize what is happening.

Our calculator runs all eight formulas simultaneously and presents the range of results, so you can see how much they diverge for your specific measurements. (Katch-McArdle and Cunningham need a body fat percentage, so they appear once you enter one.) This spread itself is informative - if all formulas agree closely, you can have more confidence in the estimate. If they diverge significantly, adaptive calibration becomes more important to determine which end of the range best matches your observed trend.

2023 DRI EER (Default for Adults 19+)

The Estimated Energy Requirement (EER) equations in the 2023 Dietary Reference Intakes come from the National Academies of Sciences, Engineering, and Medicine. Unlike the BMR formulas below, they predict total daily energy expenditure directly. They were fitted to a large database of doubly labeled water measurements, which track energy use in free-living people over one to two weeks (National Academies of Sciences, Engineering, and Medicine, Dietary Reference Intakes for Energy, 2023).

How it works: separate equations for men and women use age, height and weight, with a different set of coefficients for each physical-activity band (inactive, low active, active, very active). The calculator maps its five activity levels onto those bands.

Best for: adults 19 and over who want a TDEE estimate built from measured total expenditure rather than a resting rate times a multiplier. It is the calculator's default for that group. Younger users get Mifflin-St Jeor instead.

Mifflin-St Jeor (1990)

Developed by Dr. Mark Mifflin and colleagues at the University of Nevada and published in the American Journal of Clinical Nutrition, the Mifflin-St Jeor equation is currently the most recommended formula for predicting resting metabolic rate in non-obese adults (Mifflin MD et al., Am J Clin Nutr, 1990).

Accuracy: A systematic review by Frankenfield and colleagues found Mifflin-St Jeor predicted measured RMR within 10% in 82% of non-obese subjects - significantly better than the original Harris-Benedict (73%) - and it had the narrowest error range of the equations reviewed (Frankenfield D et al., J Am Diet Assoc, 2005).

Best for: Adults with typical body composition (body fat 15-30% for women, 10-20% for men). Does not require body fat measurement, making it accessible and practical. Slightly less accurate for very lean or very obese individuals.

Limitation: Like all weight-based formulas, it cannot distinguish between muscle and fat tissue. A muscular athlete and a sedentary person of identical weight, height, age, and sex will receive identical BMR estimates, despite likely having metabolic rates that differ by 150-300 kcal/day.

Harris-Benedict (1919, Revised 1984)

The Harris-Benedict equation, originally published in 1919 by James Harris and Francis Benedict, was derived from measurements on a relatively small sample of healthy young men and women. It remained the dominant predictive equation for over 60 years before being revised by Roza and Shizgal in 1984 to correct for systematic overestimation in certain populations (Harris JA & Benedict FG, 1919; Roza AM & Shizgal HM, Am J Clin Nutr, 1984). The calculator runs both versions.

Accuracy: Validation reviews mostly test the original equation, which lands within 10% of measured RMR for about 73% of non-obese adults. The 1984 revision corrects part of the original's overestimation, but both versions still tend to overestimate BMR for sedentary modern adults, partly because the original validation population was more physically active than the average person today. For the same inputs, both versions come out about 2-8% above Mifflin-St Jeor for typical adults.

Best for: Clinical settings where the Harris-Benedict equation is the institutional standard, or when comparing against historical nutritional assessments that used this formula. For practical personal use, Mifflin-St Jeor is preferred.

Katch-McArdle (1996)

The Katch-McArdle formula is unique among common BMR equations in that it uses lean body mass rather than total body weight as the primary variable, directly addressing the most significant weakness of weight-based formulas (Katch V & McArdle W, Nutrition, Weight Control & Exercise, 1996).

Accuracy: When body fat percentage is measured accurately (DEXA or hydrostatic weighing), Katch-McArdle outperforms all weight-based formulas for populations with atypical body composition - especially lean athletes and individuals with obesity. The equation's accuracy is directly proportional to the accuracy of the body fat measurement.

Best for: Anyone with known body fat percentage, particularly those who are leaner or more muscular than average. Athletes, bodybuilders, and those who have had their body fat measured via reliable methods.

Limitation: Entirely dependent on body fat percentage accuracy. Consumer-grade bioelectrical impedance (most bathroom scales) has error of 3-8 percentage points, which can introduce more error than it removes. Recommend using only if body fat was measured via DEXA, hydrostatic weighing, or calipers with an experienced technician.

Cunningham (1980)

The Cunningham equation comes from a 1980 reanalysis of resting-rate data from healthy adults, which found that lean body mass alone predicted resting metabolic rate well (Cunningham JJ, Am J Clin Nutr, 1980). It was not built on athletes. Its higher intercept produces systematically higher BMR estimates than Katch-McArdle for the same lean body mass, which is why sports dietitians often use it for lean, highly trained athletes, whose resting metabolic rate tends to run high.

Best for: Competitive athletes, particularly those in endurance or strength sports with training volumes of 10+ hours per week. Recreational gym-goers who exercise 3-5 times per week will usually overestimate their BMR with this formula; Katch-McArdle fits them better.

WHO/Schofield (1985)

The World Health Organization's technical report series adopted the Schofield equations as the standard reference for global nutritional assessment. These equations use age-specific coefficient brackets and were validated on a large, internationally diverse dataset, making them particularly robust for populations outside North America and Europe, and for older adults where age-related metabolic changes are more pronounced (Schofield WN, Hum Nutr Clin Nutr, 1985; WHO Technical Report Series 724, 1985).

Schofield published the equations in megajoules a day. The calculator multiplies them by 239 to get kcal, which gives these versions:

Best for: International populations, older adults (60+), clinical dietetics, and public health contexts. Also provides more appropriate estimates for individuals with very high or low body weight, where age-bracket coefficients smooth out the linear weight relationship assumed by other formulas.

Owen (1986-1987)

Oliver Owen and colleagues published separate resting-rate equations for women (1986) and men (1987) in the American Journal of Clinical Nutrition. They need only body weight (Owen OE et al., Am J Clin Nutr, 1986 and 1987).

Best for: A quick cross-check when you only know body weight. Frankenfield's review included Owen among the four most-used equations and found Mifflin-St Jeor more reliable, so treat Owen as a second opinion rather than your main estimate.

Which Formula Should I Use?

Use this decision framework to select the most appropriate formula for your situation:

Accuracy Comparison: What Research Shows

Direct comparisons between formulas using indirect calorimetry as the gold standard consistently show that no single formula dominates across all populations. Key findings from the research literature:

• Of the four most-used equations (Harris-Benedict, Mifflin-St Jeor, Owen and WHO/FAO/UNU), Mifflin-St Jeor came within 10% of measured RMR in more non-obese and obese adults than any other, and it had the narrowest error range (Frankenfield et al., J Am Diet Assoc, 2005).
• In a study of obese subjects, all common equations overestimated measured RMR, with Mifflin performing best but still averaging +117 kcal/day above measured values (Amirkalali et al., Nutr Clin Pract, 2008).
• Katch-McArdle tends to beat weight-based formulas for lean, muscular people when body fat is measured well, because it works from lean mass instead of total weight. How much it helps depends on how good the body fat number is.
• Some studies in athletes find Cunningham among the closer predictors, particularly in lean, highly trained endurance athletes, whose resting metabolic rate runs above what weight-based formulas predict.
• Between-individual variability means that even the best-fitting population-level formula will be off by more than 10% for a meaningful minority of individuals. Adaptive calibration is the best next step for improving the estimate with individual data.

The Practical Takeaway

Formula accuracy matters, but it is the starting point, not the end goal. Even with the optimal formula for your situation, real-world metabolism involves individual variation, hormonal cycling, stress, sleep, and metabolic adaptation that no static equation can capture. Use the formula estimate as a starting hypothesis, then verify it against 2-4 weeks of accurate food logging and weight data. If your weight trend does not match the predicted surplus, deficit, or maintenance calories, use adaptive TDEE calibration to refine your personal estimate.