Data Driven Fitness

Most fitness dashboards plot the number and declare victory. This one interrogates the instrument first.

Body fat fell 21.7% → 15.7% in 30 days. That puts me below the 20th-percentile body-fat cut for men 20–29, leaner than about 80% of the age group (dim_bodyfat_percentile, NHANES-derived). The fat loss is real, and it reconciles with energy balance. The lean "gain" is the part I'd be careful with.

§ 0.1 — The lean gain, decomposed

The reading shows + 8.40 lb lean. At most 4.29 lb of that is plausible newbie muscle in a month. The rest, 4.11 lb ( 4,900% %), about 0.560 gallons of water, fits training-onset glycogen loading plus a scan-day hydration difference. DEXA counts all non-bone, non-fat mass as "lean," water included.

Variables on this page

• Body fat % — fat mass as a share of total mass, measured by DEXA.
• Lean Δ / Fat Δ — change in lean / fat mass between the two scans, in lbs.
• real muscle / water — the decomposition: min(lean Δ, 4.3 lb) is the plausible newbie-muscle ceiling; the remainder is expressed as water at BodySpec's 7.4 lb-per-gallon coefficient.

What this does and doesn't show

It shows the fat loss is real and that the lean gain is roughly half non-muscle water. It doesn't claim an exact muscle figure: 4.3 lb is an upper bound, not a point estimate. And "resolvable" only means the change beats random precision noise, which isn't the same as it being muscle. Hydration is a separate and larger systematic effect. Definitions live in the data dictionary; the derivation is in methods.

§ 0.2 — Read on

• The reconciliation: what energy balance can and can't explain.
• Regional hypertrophy: did the muscles I trained most gain the most?
• Recovery: did training load move sleep, HRV, or resting HR?