Why two TDEE calculators hand you different numbers

The 1919 equation still running inside modern calculators

Harris and Benedict measured basal metabolism with a respiration chamber, one subject at a time, and reduced the results to a linear fit on four inputs: weight, height, age, and sex. For men the original reads BMR = 66.47 + 13.75 × kg + 5.003 × cm - 6.755 × age, and for women the constants shift. It was a real achievement for its era, and for sixty years it stayed the default any textbook reached for. The trouble is the sample. Those 239 subjects were mostly lean young adults from the 1910s, and the fit runs high for the heavier, more sedentary bodies common a century later. Feed it a modern desk worker and it tends to overestimate resting burn by a hundred calories or more.

One number, fitted once, propagated for a hundred years.

What each revision fixed, from Harris-Benedict to Mifflin-St Jeor

In 1984 Carla Roza and Harry Shizgal reran the arithmetic. Their paper in the American Journal of Clinical Nutrition pooled a larger dataset, roughly 337 subjects, and republished the Harris-Benedict equation with corrected coefficients, because the 1919 constants had been carried through decades of textbooks with small transcription errors. The revision moved the men's constant term from 66.47 up to 88.36 and adjusted the rest.

Six years later, Mifflin and St Jeor took a cleaner path. Rather than patch the old coefficients, they fitted a new equation on 498 modern subjects and deliberately simplified the shape:

Round numbers and a single sex offset, with nothing else separating the two cases. A 2005 review in the Journal of the American Dietetic Association tested the common equations against measured metabolism and found Mifflin-St Jeor the most reliable across both normal-weight and obese people. That is the equation TDEEfy runs.

Here is the same reference person, an 80 kg (176 lb), 180 cm, 30-year-old man, put through three generations of the fit:

The 84-calorie spread between the oldest and newest row is not measurement error. It is two different equations, fitted on two different populations seventy years apart, each doing exactly what it was built to do. TDEEfy uses Mifflin-St Jeor rather than the Harris-Benedict line because the 2005 review, and spot checks against published metabolic-cart figures, kept landing closer to it for present-day body types.

Katch-McArdle and the road through lean body mass

A fourth equation ignored the crowd. Katch and McArdle, working from exercise-physiology data, argued that fat tissue burns almost nothing at rest, so a formula keyed to total weight misreads two people who weigh the same but carry different muscle. Their equation drops height, age, and sex entirely:

For a lean, muscular subject it often lands above what Mifflin-St Jeor predicts, and for someone carrying more fat at the same scale weight it lands below. The catch is the input. You need a body-fat percentage to compute lean mass, and most people are guessing at that figure, which quietly reintroduces the error the equation was meant to remove. TDEEfy stays on Mifflin-St Jeor for that reason: four inputs anyone can supply without a caliper or a scan.

The activity multiplier was bolted on decades later

Every equation above stops at basal metabolism, the 60 to 70% of daily energy your body spends at complete rest. Total burn needs a second step, and that step has a shorter, messier history. The physical activity level concept was formalized in a 1985 WHO/FAO/UNU expert report, which expressed daily energy as a multiple of BMR. The specific ladder most fitness sites use, from 1.2 for sedentary up to 1.9 for hard physical work, is a later rounding of that idea into five convenient rungs. TDEEfy uses the standard set: 1.2, 1.375, 1.55, 1.725, and 1.9.

That second step is where most of the disagreement between calculators actually lives, more than the choice of BMR equation. Two tools can run identical Mifflin-St Jeor math and still hand you numbers 300 calories apart because one folds your workouts into the multiplier while the other asks you to log them separately, then stacks them on top. Apps such as MyFitnessPal.com lean toward the log-separately model, which double-counts if you have already picked an active multiplier. Choose the multiplier that matches a normal week, count exercise once, and the two approaches converge.

What TDEEfy runs, and the assumptions inside the number

TDEEfy computes BMR with Mifflin-St Jeor, multiplies by the activity factor you pick, and shows the result. It converts pounds and feet or inches to metric before the math runs, because the equation is defined in kilograms and centimeters. It does not read body composition, it does not know your thyroid, and it assumes your self-reported activity is honest, which is the single input people most often inflate.

None of the four equations is exact. Each is a regression fitted to a sample of real people, and the right way to use any of them is as a starting estimate you correct against the scale over two to three weeks. If the figure here disagrees with the last calculator you tried, now you can name the reason: a different equation, a different multiplier convention, or a workout counted twice.

Frequently asked

Where to go after your TDEE

Macra

Splits the calorie number you just calculated into protein, carb, and fat grams in about thirty seconds.

Tapelean

Estimates body fat from neck, waist, and height with a cloth tape, the same kind of no-lab regression this page just walked through.

Plateside

Works out which plates go on each side of the bar for any target weight, and reads back what is already loaded.