Metabolic Calibration

How NutriJourney's adaptive TDEE engine learns from your body and calibrates your targets.

Most fitness apps calculate your Total Daily Energy Expenditure (TDEE) once using a formula and never update it. NutriJourney takes a fundamentally different approach -- it treats your initial TDEE as a starting hypothesis and continuously calibrates it against your real-world results.

Why Static TDEE Fails

Standard TDEE formulas like Harris-Benedict and Mifflin-St Jeor estimate your calorie expenditure based on height, weight, age, sex, and an activity multiplier. The problem:

Activity multipliers are subjective and imprecise
Metabolic rate varies significantly between individuals with identical stats
Your metabolism adapts over time as you diet, train, or change habits
Stress, sleep, hormones, and medication all affect energy expenditure

A formula gives you a starting point. NutriJourney gives you a moving target that converges on your actual metabolic rate.

The Calibration Loop

Every week, NutriJourney runs a calibration cycle:

Predicted expenditure -- Your current calibrated TDEE multiplied by 7 days
Actual intake -- Your total logged calorie intake for the week
Expected weight change -- Based on the difference between expenditure and intake (approximately 3,500 calories per pound)
Actual weight change -- Your trust-weighted average weight at the end of the week versus the start
Discrepancy -- The gap between expected and actual weight change reveals how far off your TDEE estimate is
Adjustment -- A damped correction is applied to your TDEE for the following week

This cycle repeats every week, and your TDEE estimate gets more accurate with each iteration.

Damped Nudge Formula

NutriJourney does not overcorrect. The calibration uses a damped nudge approach:

Small corrections are applied each week rather than large swings
The adjustment is proportional to the discrepancy but capped to prevent overreaction
Over several weeks, these small nudges converge on your true TDEE
If your lifestyle changes (new job, new training program), the system adapts within 3-4 weeks

This prevents the yo-yo effect that happens when apps make dramatic calorie adjustments based on a single week of data.

Trigger Correlation Engine

After accumulating 14 or more days of data, NutriJourney runs linear regression analysis to identify how specific factors correlate with your weight fluctuations:

Carbohydrate intake -- Quantifies exactly how much water weight you retain per gram of excess carbs
Sodium intake -- Measures your body's specific sodium-to-water-retention relationship
Workout intensity -- Captures post-workout inflammation and glycogen replenishment effects
Sleep quality -- Identifies how poor sleep correlates with water retention and cortisol-driven weight changes

These correlations are unique to you. Some people retain 3 pounds of water after a high-carb day; others retain 1 pound. The trigger correlation engine learns your specific responses.

Trust-Weighted Averages

The calibration loop depends on accurate weight data, which is why trust-weighted averages are critical:

Morning fasted measurements contribute more to the weekly average than post-meal readings
Consistent measurement conditions (same time, same scale, same clothing) produce higher trust scores
Outlier readings are automatically down-weighted rather than discarded
The result is a weekly average that represents your true weight more accurately than any single reading

What You Will See

The calibration engine works in the background, but its results are visible throughout the app:

Dashboard -- Your calibrated TDEE displayed prominently, along with a confidence indicator that grows as calibration matures
My Body tab -- Trigger correlation insights explaining recent weight movements, trust-weighted trend lines, and noise band visualization
JUNA coaching -- JUNA references your calibrated TDEE and trigger data when discussing calorie targets, explaining weight changes, or suggesting adjustments
Macro targets -- Your daily protein, carb, and fat targets are derived from your calibrated TDEE and automatically update as calibration refines the estimate