Original analysis
Where the Standard Fitness Formulas Disagree
4 analyses that take the equations behind every fitness calculator on the internet, run them across systematic ranges of adult bodies, and report where they contradict each other and by how much. Every number is computed when this site is built, from the same code that powers the calculators, so all of it can be checked.
Why we publish this
A calculator gives you one number and, usually, a great deal of confidence about it. Behind that number are equations fitted decades ago to particular populations, and there is almost always more than one of them. Two respectable formulas can be built into two respectable calculators and hand the same person answers hundreds of calories apart. Nobody publishes how far apart, for whom, or whether it matters next to the other choices you make when filling in the form.
These pages answer those questions the only way they can be answered without new clinical data: by running the formulas exhaustively and reporting what came back. Three of the four analyses do that by sweep — 2,702 computed body-by-formula combinations so far. The fourth needs no sweep, because the result falls out of the equations exactly. None of this is a substitute for measured research, and none of it establishes what anyone should eat or weigh. It establishes what these equations actually assert, which turns out to be less settled than the tools built on them suggest.
How each page is structured
The same order every time: the question, then what we found, then how it was computed, then the analysis, then what would make it wrong, then how to reproduce it. The limitations section is not a disclaimer — it is where the honest boundaries of each result live, including the cases where we decided a figure could not be produced responsibly and left it out.
Two consequences of doing this seriously. First, several findings are unflattering about tools on this site, and they are published anyway. Second, where an analysis found that our own copy overstated something or omitted a caveat, we changed the copy — each page says which change it caused, and those changes are live. Analyses are written and reviewed by Jesse, and our methodology page documents every formula the site uses.
The analyses
How Much Do BMR Formulas Actually Disagree?
Mifflin-St Jeor and Harris-Benedict are the two formulas almost every calorie calculator uses. How far apart do they land, for whom, and does the choice matter more or less than picking your activity level?
What we found: Median gap between the two formulas: +57 kcal/day (3.8%), rising to 270 kcal/day in the worst case. Choosing the wrong activity level is 2.9 times larger an error, in 98% of cases.
Last reviewed August 20, 2026
The Ideal Weight Formulas Contradict Each Other by Height
The four classic ideal-body-weight formulas are all linear in height, but BMI is quadratic. What BMI does each formula actually imply, and what happens to their advice at the ends of the height range?
What we found: The four formulas disagree by up to 16 kg at the same height, and they split evenly on direction: 4 of the 8 formula-by-sex combinations imply a rising BMI as height increases, 4 a falling one. The healthy BMI range is about 3.8 times wider than the whole argument.
Last reviewed August 20, 2026
How Wrong Is the 7,700-Calorie Weight Loss Rule After a Year?
Every weight-loss planner, including ours, assumes a fixed daily deficit produces a fixed weekly loss. Your resting metabolism falls as you lose weight, so it cannot. How large is the error after six and twelve months?
What we found: After a year on a fixed calorie target, the straight-line projection overshoots by a median of 26.1% of the predicted loss — and only 53% of the original deficit is still doing any work by then.
Last reviewed August 20, 2026
BMI and Body Fat Percentage Cannot Be Reconciled
BMI ignores every circumference; the U.S. Navy body-fat method ignores body weight entirely. What happens when two methods built on disjoint inputs are asked to classify the same person?
What we found: The two methods share 1 of the 5 measurements they read between them. Hold the tape fixed and body weight can cross 3 BMI categories without moving the body-fat estimate at all.
Last reviewed August 20, 2026
Checking any of this
Each analysis names the functions it called, and every one of those functions is reachable through a calculator or a chart on this site, so single rows can be verified by hand. The guides cover what to do with the results in practice. If your own rebuild disagrees with a figure here, the contact page reaches a person who will check it and correct the page if it is wrong.