Why TDEE Calculators Disagree by 800 Calories — and How to Fix It

An 8-minute read on the math behind daily energy expenditure, why every online calculator gives you a different number, and how to break the math apart to land on something real.

---

TL;DR

Online TDEE calculators routinely disagree by 600–900 kcal for the same person. The reason is almost never the BMR math (everyone uses Mifflin-St Jeor) — it's the single "activity level" dropdown that squashes all your non-sleeping life into five crude buckets.

The fix is to stop using the multiplier and decompose your burn into its actual four components:

TDEE = BMR + NEAT + TEF + EAT

Estimate each one from your actual life. You land on a tighter number, and — just as important — you can see which piece is wrong when the scale argues with you later.

Want to skip the math? Our Accurate TDEE Calculator runs this decomposition for you in five questions — no signup, no email. The rest of this post explains what's happening under the hood.

---

The problem: the activity multiplier is a 5-bucket lie

Plug your stats into eight different calculators and you'll get eight different TDEE numbers. Not slightly different — meaningfully different. I ran my own (28M, 175 cm, 80 kg) through a batch and got results from ~1,800 to ~2,600 kcal.

Here's why. Every calculator does this:

1. Compute BMR (basal metabolic rate) with the Mifflin-St Jeor equation. All eight agreed to within 5%.
2. Ask you to pick an "activity level" from a dropdown:
   • Sedentary (× 1.2)
   • Lightly active (× 1.375)
   • Moderately active (× 1.55)
   • Very active (× 1.725)
   • Extra active (× 1.9)
3. Multiply.

Step 2 is the problem. The spread between "lightly" and "moderately" active, on a 1,759 kcal BMR, is:

1,759 × 1.375 = 2,419
1,759 × 1.55  = 2,726
Difference     = 307 kcal/day

That's an entire intentional deficit, hinging on a dropdown I can't clearly answer. Am I "lightly" active because I drive everywhere and work from home? Or "moderately" because I lift four times a week? Different calculators define these buckets differently. Some count exercise, some don't. None show their work.

Pick the wrong bucket and you'll either not lose weight ("but I'm eating the deficit the calculator told me!") or lose too fast and crash. It's the root cause of the most common frustration in weight-loss communities: "CICO isn't working."

CICO works. The calculator's guess at your O is just bad.

---

The fix: decompose

Your body burns energy in exactly four categories. Traditional calculators jam three of them into one multiplier. Split them back out:

Component	What it is	Rough % of total
BMR	Basal metabolic rate — what you burn existing	60–70%
NEAT	Non-Exercise Activity Thermogenesis — all non-workout movement	15–30%
TEF	Thermic Effect of Food — energy cost of digestion	8–12%
EAT	Exercise Activity Thermogenesis — actual workouts	5–15%

Estimate each one independently. The dropdown disappears.

---

Component 1: BMR

The boring, reliable part.

Mifflin-St Jeor (most accurate for the general adult population):

• Male: BMR = 10 × weight_kg + 6.25 × height_cm − 5 × age + 5
• Female: BMR = 10 × weight_kg + 6.25 × height_cm − 5 × age − 161

For 28M, 175 cm, 80 kg:

10 × 80 + 6.25 × 175 − 5 × 28 + 5
= 800 + 1,093.75 − 140 + 5
= 1,759 kcal/day

If you know your body fat percentage (DEXA scan, BodPod, or a decent smart scale corrected with calipers), Katch-McArdle is more accurate because it works off lean body mass:

BMR = 370 + 21.6 × lean_body_mass_kg

Accuracy on both: ±10% for about 80% of the population. If you're an outlier (very muscular, very old, certain medical conditions), no formula will be great — but the component-decomposition approach at least lets you correct the pieces downstream.

Typical range: 1,200 – 2,000 kcal/day.

---

Component 2: NEAT — the one that matters most

NEAT is everything your body does that isn't sleeping and isn't a deliberate workout. Walking to the kitchen. Standing at a counter chopping vegetables. Fidgeting in a meeting. Carrying groceries. A restless person might burn 900 kcal/day here. A desk worker who orders delivery might burn 200. That 700 kcal delta is exactly why traditional calculators fail — it's a factor of 4, and the dropdown can't see it.

Estimate NEAT by scoring four sub-dimensions.

A. Occupation (biggest contributor)

Job type	Examples	kcal/day
Desk, remote	WFH, minimal movement	200
Desk, office	Office with walks to meetings, coffee, etc.	250
Standing, light	Retail, reception, lab work	400
Standing, active	Teaching, nursing, restaurant server	550
Physical, light	Warehouse, postal delivery	650
Physical, heavy	Construction, farming, moving	800

B. Daily steps (if you track them)

Steps/day	kcal
< 3,000	75
3,000 – 5,000	125
5,000 – 7,500	200
7,500 – 10,000	275
10,000 – 12,500	350
12,500 – 15,000	425
15,000+	500

C. Commute (round trip)

Method	kcal
Car / taxi	0
Public transit (with walking to/from)	50 – 100
Bike (per 30 min)	100 – 150
Walk (per 30 min)	80 – 120

D. Household & caregiving

Activity	kcal
Minimal (takeout, no chores)	0
Cook daily	50 – 80
Regular cleaning / laundry	30 – 60
Childcare (young kids)	100 – 200

Summing NEAT without double-counting

Occupation and steps overlap (your desk job's 200 kcal already includes some walking around the office). If you have step data, use the higher of the two as the "base movement" number, then add commute and household on top:

NEAT = max(occupation, steps) + commute + household

If you don't have step data, just use the occupation estimate.

Worked example (me): desk/remote (200) + drive commute (0) + cook dinner most days (50) = 250 kcal NEAT. A friend with the same BMR who's a nurse pulling 12-hour shifts + hospital walking: 550 + ~40 for light cleaning = 590 kcal. Same BMR, 340 kcal daily difference, from NEAT alone. No dropdown can capture that.

Typical range: 200 – 900 kcal/day.

---

Component 3: TEF

Digesting food costs energy. Different macronutrients cost different amounts.

Quick estimate (accurate enough for mixed diets):

TEF = total_daily_intake × 0.10

Precise estimate (when you know your macros):

Macro	TEF rate	kcal cost per 100 g
Protein	25%	100
Carbs	7.5%	30
Fat	1.5%	13.5

TEF = (protein_g × 4 × 0.25) + (carb_g × 4 × 0.075) + (fat_g × 9 × 0.015)

This is where TEF stops being a rounding error and starts being interesting.

Same 2,400 kcal eaten three different ways:

Diet	Macro split (P/C/F)	TEF
Keto	25 / 5 / 70	~170
Standard western	15 / 55 / 30	~230
High protein	35 / 35 / 30	~290

That's a 120 kcal/day spread from the same total calories. Partly explains why high-protein diets "feel" more effective — you literally burn more energy processing the food. No traditional TDEE calculator accounts for this.

Typical range: 150 – 300 kcal/day.

---

Component 4: EAT

Calories burned during deliberate exercise. Use MET values (Metabolic Equivalent of Task) — a dimensionless multiple of resting metabolism, measured in the lab for basically every activity humans do.

The formula:

EAT_per_session = (MET − 1) × weight_kg × duration_hours
EAT_daily_avg   = EAT_per_session × sessions_per_week / 7

Why (MET − 1) and not just MET? MET is defined inclusive of your resting metabolism (1 MET = your BMR burn rate). If you use the raw MET number you'll double-count the BMR you already added at the top. Subtract the 1 to isolate the additional burn from moving.

MET values for common exercise:

Activity	MET
Yoga (gentle)	2.5
Walking, casual (3 mph)	3.5
Weight training, moderate	3.5
Walking, brisk (4 mph)	5.0
Elliptical	5.0
Dance	5.5
Weight training, vigorous	6.0
Hiking	6.0
Swimming, moderate	7.0
Rowing machine	7.0
Cycling, moderate	8.0
Jogging (5 mph)	8.0
HIIT / CrossFit	8.0
Basketball, soccer	8.0
Jump rope	10.0
Running (7 mph)	11.0

Worked example: 80 kg, weight training vigorously, 60 min × 4 sessions/week.

Per session: (6 − 1) × 80 × 1.0 = 400 kcal
Daily avg:   400 × 4 / 7 ≈ 229 kcal/day

Two things to notice. First: across the whole week, four solid lifting sessions come out to ~230 kcal/day. That's less than most people assume. Second: do not eat your exercise calories back. The MET-based number is an average over the week and is already contributing to your maintenance budget. "Burn 500, eat 500 more" is how you erase your deficit.

Typical range: 0 – 400 kcal/day (averaged daily).

---

Worked example: full breakdown

28 year old male, 175 cm, 80 kg, software engineer working from home, drives everywhere, ~5,000 steps on rest days, lifts vigorously 4×/week for an hour, eats a high-protein diet (35/35/30).

────────────────────────────────────────────────
  BMR   Mifflin-St Jeor                1,759
        10×80 + 6.25×175 − 5×28 + 5

  NEAT  Desk/remote job                  200
        Drive commute                      0
        Cook dinner most days             50
        (Steps 5,000 ≈ 125, lower than
         desk estimate, so we ignore)
                              NEAT =    250

  TEF   High protein (35/35/30)
        on ~2,500 kcal intake           ~290

  EAT   Weight training, MET 6
        60 min × 4/week
        (6−1) × 80 × 1.0 = 400/session
        400 × 4/7 ≈                     229
────────────────────────────────────────────────
  TDEE                                2,528 kcal
────────────────────────────────────────────────

  Traditional comparison:
  "Lightly active" × 1.375  →  2,419  (off by  −109)
  "Moderately active" × 1.55 →  2,726  (off by  +198)

The real number sits right between the two buckets the dropdown would have forced on me. If I'd picked "lightly active" I would have been surprised I wasn't losing weight. If I'd picked "moderately active" I would have cut 300 kcal too aggressively.

---

How to use this on yourself — 6 steps

1. Compute your BMR. Plug your stats into Mifflin-St Jeor. If you know body fat %, compute Katch-McArdle too and use whichever you trust more.
2. Estimate NEAT. Pick your occupation row. Add commute. Add household. If you have step data from a watch, use max(occupation, steps) as the base.
3. Estimate TEF. If you don't know your macros yet, use 10% of your target intake. If you do, use the macro formula.
4. Estimate EAT. For each exercise you do regularly, compute (MET − 1) × kg × hours per session, multiply by sessions/week, divide by 7.
5. Sum them. This is your starting TDEE estimate.
6. Validate against reality. Track your intake and weight for 2–3 weeks. If you eat TDEE − 500 and lose ~0.5 kg/week, you nailed it. If not, don't panic — this is where decomposition wins. Look at the pieces. Did you overestimate NEAT (most common)? Did you eat back exercise calories without meaning to? Is your step count actually lower than you think? Each component is individually auditable.

That last step is the whole point. Traditional calculators give you a single number and no explanation when it's wrong. Decomposed estimates give you a breakdown you can correct piece by piece.

---

FAQ

Isn't this over-engineered? Can't I just track and adjust? Sure, you can skip to step 6 and run the experiment directly. But you still need a starting point, and the traditional dropdown gives you a starting point that's wrong by hundreds of calories. Decomposition narrows the initial error so the experiment converges faster — usually 2–3 weeks instead of 2–3 months.

What about metabolic adaptation? Real but overstated. In most people, metabolism slows by 50–100 kcal/day during an aggressive cut (not the 500+ that "starvation mode" myths claim). Recalculate every 5 kg of weight loss — BMR drops by roughly 20 kcal per kg lost, and NEAT tends to drop as you're less bouncy on a deficit. Plug the new weight into the equations and re-derive.

What if I don't have a step counter? Use the occupation row alone and skip the step dimension. Your estimate will be less precise but still better than the traditional multiplier, because you're at least isolating commute and household separately.

How do I recalculate when I lose weight? Every 5 kg. BMR re-computes directly from the new weight. NEAT's occupation/commute/household numbers don't change, but EAT does — (MET − 1) × new_weight × hours — which slightly drops your exercise burn. The system's not static; weight loss reduces your TDEE by ~30 kcal per kg lost. Plan for it.

Where do the MET numbers come from? The Compendium of Physical Activities (Ainsworth et al., updated 2011 and 2024), which is the source academic researchers use. These are measured averages; individual variation exists but MET-based estimates are far more accurate than "moderate intensity × 3 hours/week" treatments.

Why doesn't MyFitnessPal do this? Because an activity dropdown is a one-click onboarding step and a four-dimensional NEAT breakdown is not. The existing apps were designed for a different bar on accuracy vs. friction. The trade-off is worth revisiting now that we have tools (conversational UIs, step data from phones, wrist-worn trackers) that make the four-dimensional input feel as lightweight as one dropdown.

---

Closing note

We turned this decomposition into a free tool, the NanoRhino Accurate TDEE Calculator. It walks you through the same four components as a quick conversation — five or six questions about your life, then a labeled breakdown of where your calories actually go. Every number is challengeable, every assumption is shown. No signup, no email, no app to install.

If you'd rather do it on paper, the math in this post is enough. Either way, I'd love to hear how close the estimate lands to your real maintenance number once you've tracked a few weeks of intake and weight.

For ongoing nutrition coaching once you have your number, text START to (915) 277-7888 — your AI nutrition coach takes it from there.

Built with care by Link Heart Limited in Houston, Texas.