Methodology & Corrections

Model version 1.0.0 · Effective 24 July 2026

1. How to read the ranges & confidence

Where BodEvo shows a range like 1,900–2,300 kcal, the middle is the model's point estimate and the width reflects normal variation and input uncertainty — it is not a guarantee that reality falls inside it. Confidence is a plain three-level indicator:

Rule of thumb: the direction and relative size of a change are more reliable than any single decimal place.

2. Maintenance / expenditure calories

Formula. Basal metabolic rate (BMR) uses one of two established equations:

• Mifflin-St Jeor (default): BMR = 10·kg + 6.25·cm − 5·age + s where s = +5 (male) or −161 (female).
• Katch-McArdle (when body-fat % is provided): BMR = 370 + 21.6·LBM(kg), where lean body mass LBM = weight·(1 − bodyfat%).

Maintenance = BMR × an activity multiplier (1.2 sedentary → ~1.9 extremely active), or, in detailed mode, BMR plus separately-estimated job, step, lifting, cardio, sports, and thermic-effect components (each a MET-based population estimate).

Limits. Population equations carry roughly ±10% individual error even with perfect inputs; activity multipliers are coarse buckets; the component breakdown multiplies several estimates together. That is why the static calculator figure is shown as a band. Once you log food and weigh-ins, the app measures your real expenditure from energy balance (see the adaptive expenditure card in the Daily Tracker) and confidence rises with data.

Sources: Mifflin MD et al., Am J Clin Nutr 1990; Katch & McArdle; Compendium of Physical Activities (MET values).

3. Body-fat percentage

Body-fat % is a value you enter (from calipers, DEXA, smart scale, or a visual estimate). BodEvo does not measure it. Where the projection shows body-fat % or fat-/lean-mass changing over future weeks, that is derived from your starting value plus the partitioning heuristic below — so its accuracy is capped by the accuracy of the number you entered.

Limits. Consumer body-fat methods commonly disagree by 3–5 percentage points. Treat projected body-fat % as a trend indicator, not a caliper reading.

4. Weekly weight change & the projection curve

Formula. The baseline weekly change uses the standard energy-balance approximation: weekly lb ≈ (daily calorie balance × 7) ÷ 3,500 (7,700 kcal per kg). The multi-week curve iterates this, and each week it:

• recomputes maintenance as body weight changes,
• applies a metabolic-adaptation factor as a deficit/surplus persists (maintenance drifts down on a cut, up on a bulk),
• applies a plateau factor so loss/gain slows realistically rather than continuing in a straight line.

Limits. The 3,500-kcal-per-pound rule is a planning heuristic, not a biological constant; water and glycogen swings dominate any single week; logging and weighing accuracy set a floor on precision. The projection is a plan, not a prediction — the weekly check-in re-fits it to your measured trend.

5. Fat vs. lean partitioning (muscle-gain / fat-gain heuristics)

How it works. A base fat/lean fraction (e.g. ~90% of loss as fat on a cut, ~35–50% of gain as muscle on a bulk) is nudged by inputs with established directional effects: deficit/surplus size, protein per lb, sleep, training frequency, current body-fat %, and the partitioning score. The fraction is clamped to a plausible window, then multiplied by the weekly change to give fat- and lean-mass ranges (±8% on a cut, ±12% on a bulk). "Fat-gain risk" and the confidence level are threshold labels over those same inputs.

Limits. Actual partitioning depends on genetics, training history, drugs, and measurement method, none of which this captures precisely. Use it to compare scenarios (e.g. more protein vs. less), not as a body-composition readout.

6. Strength projections

Formula. Projected strength applies an experience-scaled weekly progression rate (beginners progress fastest, advanced lifters slowest) to each lift, modulated by the nutrition context (calorie balance, protein, sleep, recovery/fatigue) and tapered over time so the curve flattens rather than growing without bound.

Limits. Individual strength progression varies enormously with programming, technique, sleep, and genetics. Projections assume consistent, well-run training and no injury or long layoff. Treat them as a motivating best-case trajectory, shown as a range that widens the further out it goes.

7. Composite 0–100 scores (stress, recovery, partitioning)

Scores such as training stress, recovery, fatigue, and the nutrient-partitioning score are BodEvo's own 0–100 composites: weighted combinations of your inputs (volume, frequency, sleep, calorie balance, protein, body-fat) mapped onto a 0–100 scale. They are relative indicators for comparing scenarios and tracking change over time, not measured physiological quantities and not comparable to any lab metric or another app's score.

Limits. The weightings are chosen heuristics. A "72/100" is meaningful only against your own other values, not as an absolute.

8. Corrections & changelog

Formula changes and corrections are logged here with the model version. If a fix changes the numbers you see, it appears in this list.