BMR Calculator Results: 1,395 vs 1,485 Calories, Why?
I once ran the exact same numbers through four different calorie apps in one sitting.
Same height. Same weight. Same age.
Four different BMR numbers came back.
I remember staring at my phone thinking I'd typo'd something. I hadn't. What I'd actually stumbled into is a genuinely two-layered problem: calculator apps don't just round differently, they're often built on entirely different formulas, and on top of that, they apply a wildly inconsistent "activity multiplier" that can move your final number even more than the formula choice does.
Sound familiar? Let's break down exactly why, with real numbers.
Reason #1: There's No Single BMR Formula
Calculators disagree first because basal metabolic rate (BMR) isn't measured from one universally agreed formula — it's estimated from different regression equations, built from different study populations, sometimes decades apart.
The two you'll run into most are Harris-Benedict and Mifflin-St Jeor.
Harris-Benedict (1919, revised 1984) was originally derived from 239 subjects at the Carnegie Institution — mostly lean, young, physically active American men and women, measured by direct calorimetry. In 1984, Roza and Shizgal revised the coefficients with newer data:
- Men: BMR = 88.362 + (13.397 × weight in kg) + (4.799 × height in cm) − (5.677 × age)
- Women: BMR = 447.593 + (9.247 × weight in kg) + (3.098 × height in cm) − (4.330 × age)
Mifflin-St Jeor (1990) came later, built from 498 subjects at the University of Nevada School of Medicine, measured by indirect calorimetry — a population with more body fat on average than the 1919 Harris-Benedict cohort, closer to a modern adult population:
- Men: BMR = (10 × weight in kg) + (6.25 × height in cm) − (5 × age) + 5
- Women: BMR = (10 × weight in kg) + (6.25 × height in cm) − (5 × age) − 161
Because the original Harris-Benedict dataset reflects a more active, leaner, different-era population, applying it (even the 1984 revision) to a modern, more sedentary person tends to overestimate BMR by roughly 5% to 15%.
Same Person, Two Formulas: Here's the Actual Number Gap
Here's what that looks like with real people plugged in. A published comparison ran identical profiles through both formulas:
| Profile | Mifflin-St Jeor | Harris-Benedict (revised) | Difference |
|---|---|---|---|
| 35-year-old woman, 5'5", 150 lb | 1,395 kcal | 1,485 kcal | +90 kcal |
| 35-year-old man, 5'10", 175 lb | 1,755 kcal | 1,820 kcal | +65 kcal |
| 50-year-old man, 5'10", 220 lb | 1,910 kcal | 2,055 kcal | +145 kcal |
Notice two things. First, Harris-Benedict came out higher than Mifflin-St Jeor in every single profile. Second, the gap grows with body weight — from 65 kcal at a lighter build up to 145 kcal for a heavier profile.
So yes, formula choice alone can shift your number by roughly 65 to 145 calories. That's real, and it explains part of why your app doesn't match your friend's app, or your doctor's chart.
But it's not the biggest source of disagreement. Not even close.
The Bigger Culprit: Your Activity Multiplier
Once you have a BMR number, apps multiply it by an "activity multiplier" to estimate your Total Daily Energy Expenditure (TDEE) — and this is where the real spread happens.
The textbook scale, originally tied to the Harris-Benedict framework, looks like this:
- Sedentary (little/no exercise): × 1.2
- Lightly active (light exercise 1–3 days/week): × 1.375
- Moderately active (moderate exercise 3–5 days/week): × 1.55
- Very active (hard exercise 6–7 days/week): × 1.725
- Extra active (hard daily exercise + physical job): × 1.9
Here's the catch: not every app uses this exact scale. One comparison put it bluntly — "one site might say 'moderate exercise' is 1.55 times your BMR. Another might use 1.65." (Worth flagging: that specific comparison comes from a general wellness blog rather than a peer-reviewed source, so treat the exact numbers as illustrative rather than gospel — but the underlying point, that multiplier scales genuinely differ across tools, holds up.)
FitnessVolt documents an entirely separate multiplier scheme some tools use — sedentary ×1.15, light ×1.2–1.35, moderate ×1.4–1.55, very active ×1.6–1.75, extra active ×1.8–1.95 — which overlaps with the "standard" scale but isn't identical to it. FitnessVolt states plainly that the activity-level multiplier "has a larger impact on your final calorie estimate than almost any other factor," and that picking the wrong activity level alone can shift your calorie target by 300 to 700 calories.
Compare that to the 65–145 kcal gap from formula choice above. The activity multiplier isn't a rounding error — it's very often the main event.
Wait, Why Does the Government Use a Different Formula?
Here's where it gets structurally interesting. The BMR-times-multiplier approach isn't the only model in circulation. The U.S. National Academies' Institute of Medicine (IOM) publishes its own Estimated Energy Requirement (EER) equation as part of the 2005 Dietary Reference Intakes — and it doesn't calculate BMR first and multiply afterward at all. It folds a "physical activity coefficient" (PA) directly into a single regression:
- Women: EER = 354 − (6.91 × age) + PA × [(9.36 × weight in kg) + (726 × height in m)]
- Men: EER = 662 − (9.53 × age) + PA × [(15.91 × weight in kg) + (539.6 × height in m)]
The PA value here only ranges from about 1.0 (sedentary) to roughly 1.45–1.48 (very active) — a noticeably narrower, lower-ceiling range than the 1.2–1.9 scale used by Harris-Benedict/Mifflin-based calculators. That's not just a different multiplier on the same base number. It's a structurally different equation. Which means a government-lineage tool and a commercial app can land on genuinely different TDEE numbers for the exact same person, for reasons that have nothing to do with rounding or typos.
Real-world apps aren't standardized either. Noom, one of the most widely used U.S. weight-loss apps, states it calculates total daily energy expenditure using the Harris-Benedict equation, then layers its own dynamic "Weight Loss Zone" that adjusts based on logged activity — roughly half of workout calories burned above a BMR-based threshold get added back, and each logged step adds about 0.05 calories to the daily budget. So Noom's number isn't just "Harris-Benedict plus a multiplier" — it's Harris-Benedict plus proprietary logic layered on top. MyFitnessPal, Lose It!, and other popular apps each make their own formula and multiplier choices, which is a big part of why cross-app comparisons rarely match.
Which Formula Is Actually More Accurate?
If you only read one section, make it this one: peer-reviewed evidence points fairly clearly toward Mifflin-St Jeor as the better general-purpose default.
The most-cited study here is Frankenfield, Roth-Yousey and Compher's 2005 systematic review in the Journal of the American Dietetic Association, which tested Mifflin-St Jeor, Harris-Benedict, Owen, and WHO/FAO/UNU equations against actual measured resting metabolic rate. Its conclusion, echoed across secondary literature: Mifflin-St Jeor predicted resting metabolic rate within 10% of the measured value in more people — both non-obese and obese — than any other equation tested, with the narrowest error range of the group.
A separate 2013 study in Clinical Nutrition found Mifflin-St Jeor had the smallest mean error of the equations tested — about −1 kcal in non-obese subjects and +1.5% in obese subjects — while Harris-Benedict systematically overestimated BMR in both groups, by roughly 5% to 15%.
Reported "hit rate" — the share of people whose resting metabolic rate the formula predicted within 10% accuracy — varies by study: Harris-Benedict has been reported anywhere from 38% to 80%, while Mifflin-St Jeor has been reported between 70% and 82%. That wide range for both formulas is itself telling: accuracy depends heavily on which population (age, BMI, ethnicity) the study tested, which is a big part of why "my calculator disagrees with my doctor's calculator" in the first place.
The practical takeaway that shows up consistently across sources: use Mifflin-St Jeor as your general-purpose default, and switch to Katch-McArdle only if you have a reasonably reliable body fat percentage reading.
What No Calculator Can Measure
Even the best formula is still an estimate, not a measurement. Cleveland Clinic lists what standard BMR equations simply cannot capture on an individual level: genetics, muscle mass, body fat percentage, thyroid hormone levels, illness or injury, and temporary factors like fasting, ambient temperature, and stimulant use (caffeine, nicotine). A truly accurate BMR reading requires strict lab conditions — complete rest, 12 to 14 hours after sleep — which is exactly why any formula-based number is, by definition, an estimate.
There's also evidence your own BMR isn't fixed over time. NIH/NIDDK research on "The Biggest Loser" contestants found that after an average 58.3 kg weight loss during the competition, resting metabolic rate had dropped by 610 kcal/day — a bigger decline than weight loss alone would predict. A 2022 reinterpretation by lead researcher Kevin Hall attributes this to a "constrained" model of energy expenditure: sustained, intense exercise appears to trigger compensatory metabolic slowdowns that keep total energy expenditure from rising as much as the added activity alone would suggest. In other words, even your own BMR six months from now may not match the number a calculator gives you today.
And it's not just the BMR side of the equation that's shaky. A 2026 study covered by ScienceDaily, involving NIDDK and reported via the American Society for Nutrition, found that four popular AI-assisted calorie-tracking apps — MyFitnessPal, Lose It!, Cal AI, and Appadiet — underestimated calories by an average of 250 to 345 kcal per meal, and underestimated fat by about 30 grams per meal, with accuracy getting worse for lower-calorie and higher-fat meals. That's a separate error source from BMR/TDEE estimation — it's about logging your intake, not calculating your output — but it stacks on top of everything above. Consumer calorie tools can carry meaningful, app-specific error at more than one stage of the math.
What's the Most Accurate Method, Really?
Indirect calorimetry is the closest thing to a gold standard, because it directly measures oxygen consumption and carbon dioxide production rather than inferring energy expenditure from height, weight, age, and sex. It's just rarely used outside clinical settings, because it's labor-intensive and expensive.
Even indirect calorimetry doesn't make every predictive equation obsolete for every person. One review found predictive equations correlated well with indirect calorimetry in only 40% to 60% of the studied population, and accuracy varied by BMI category — one equation (Ravussin) performed best for people with overweight, while Mifflin-St Jeor and the Henry equation performed best for people with obesity. There's no universal winner across every body type; there's a "best default" (Mifflin-St Jeor) and situational alternatives.
Frequently Asked Questions
Q. Which BMR formula should I actually trust?
A. Mifflin-St Jeor is the evidence-backed default for most people — peer-reviewed research consistently shows it predicts resting metabolic rate more accurately than Harris-Benedict across both non-obese and obese populations. If you know your body fat percentage from a reliable method, Katch-McArdle is a reasonable alternative.
Q. Why does my app's calorie target seem so different from a friend's, even with similar stats?
A. It's very likely the activity multiplier, not the BMR formula. Picking the wrong activity level can shift your target by 300 to 700 calories, according to FitnessVolt — a bigger swing than the 65–145 kcal gap between BMR formulas.
Q. Is a government or clinical calculator more accurate than a commercial app?
A. Not necessarily more accurate, just structurally different. The IOM/EER equation folds activity directly into a single regression instead of multiplying BMR afterward, using a narrower PA range (roughly 1.0–1.48) than the 1.2–1.9 scale most commercial apps use. Different structure, not automatically a better number.
Q. What's the single most accurate way to know my real BMR?
A. Indirect calorimetry, done in a clinical setting under controlled conditions (rested, fasted, calm). It's the closest thing to a direct measurement rather than an estimate, though it's not widely accessible outside medical or research settings.
So, Which Calculator Should You Actually Trust?
Stop hopping between calculators looking for the "right" number — there isn't one universal right answer, and the research backs that up. Pick Mifflin-St Jeor as your default, double-check that your activity level actually matches your real daily movement (not just your workout days), and then track your own trend over weeks, not the single number any one app spits out.
If you want to go further, this is a good spot to compare notes: which formula does your favorite app actually use, and does the activity level you picked really match how much you move in a day? That single check tends to matter more than which formula you're using in the first place.
Comments
Post a Comment