The Data Behind My Body Recomposition

How I lost fat and built muscle at the same bodyweight — and why simple, consistent habits beat chasing the perfect diet, workout, or supplement.

Key takeaway: Even though my body weight stayed nearly the same for 16 months, I became noticeably leaner and stronger. This shows that long-term consistency—not chasing the perfect diet, workout, or supplement—was the biggest driver of my results.

The result in one minute

Avg calories
3,069
Avg protein
169 g
Avg steps
18,699
Training
4–5 /wk
MetricStartEndChange
Bodyweight164.6 lb163.8 lb−0.8
Body fat14.3%13.1%−1.2 pts
Muscle mass81.3%82.4%+1.1 pts
Fat mass*23.5 lb21.5 lb−2.1 lb
Lean mass*141.1 lb142.3 lb+1.3 lb

*Directional estimates, not lab-grade.

Takeaway: a textbook recomposition — fat down, muscle up, weight flat — but modest in size, exactly as expected for an already-lean, trained individual.

The question

The experiment set out to answer one question: can a natural lifter with a physically active lifestyle build muscle while staying lean?

Yes. Body recomposition—losing fat while building muscle at the same time—is one of the most difficult goals in fitness. The challenge is physiological: fat loss usually requires an energy deficit, while muscle growth typically benefits from an energy surplus. Recomposition asks the body to perform both opposing tasks simultaneously.

For beginners or individuals carrying significant body fat, this is often achievable because they have more metabolic flexibility and untapped potential for adaptation. For someone who is already lean and resistance-trained, however, that margin for improvement becomes extremely small.

That is precisely the condition tested in this experiment.

The study and the starting point

Sixteen months of continuous tracking — January 2025 to April 2026, roughly 485 days — from four independent sources:

Nutrition — daily FatSecret food logs (~485 days), aggregated monthly: calories, protein, carbs, fat.
Body composition — 16 monthly VitaFit bioimpedance readings.
Activity — Apple Health, stored as average daily steps per month (~9.1 million steps in total over the 16 months).
Training — a logged resistance program across four quarterly phases (~4–5 sessions/week).

The one gap in this dataset is sleep tracking. I averaged about eight hours a night, but without a daily sleep log, that figure stays as context rather than analyzable data. It's also important to set expectations on resolution: this is a monthly analysis built on sixteen data points, not daily measurements.

The starting point, January 2025: 164.6 lb, 14.3% body fat, BMI 23.2 — about 141 lb of lean mass, from an experienced lifter (3–5 years training) already active at ~13,000 steps in the slowest month. Lean, trained, and active before the experiment even began — so no easy beginner gains were waiting.

Sixteen months, four phases

Weight (lb) Body fat % Muscle %
Weight flat; body fat 14.3 to 13.1%; muscle 81.3 to 82.4%.
PHASE 1 · JAN–APR 2025
Foundation
Trained a bodybuilding split through Ramadan while fasting, and reached the leanest reading of the entire study — 12.4% body fat in April.
PHASE 2 · MAY–AUG 2025
Lean gain
Lifting plus one run a week, building to a half marathon in August. Squat climbs 60 → 105 plates; weight ticks up to ~167 lb as muscle and training volume build.
PHASE 3 · SEP–DEC 2025
Plateau
Walking around 20,000 steps a day with 3–4 workouts a week, while eating more on a whole-food diet — calories run highest here (3,300+/day). Weight and strength plateau near their peak.
PHASE 4 · JAN–APR 2026
Optimization
~3 runs/week building toward a May 2026 marathon, plus upper-body lifting and a deliberate higher-fat/lower-carb diet. Ends leanest (13.1%) with muscle at its highest (82.4%).
Takeaway: the gains weren't linear. A foundation, a build, a stall, and a re-optimization — with the leanest finish coming only after a deliberate change.

The variables I expected to matter most barely moved the needle

The obvious approach is to line up inputs against outcomes and look for the relationship. I ran those correlations — and they're a lesson in why single-subject, monthly data demands humility. Each chart below answers a specific question.

Did eating more mean weighing more?

Calories vs weight: no relationship, r = +0.05.
Monthly calories vs bodyweight. Correlation r = +0.05 — essentially none.

No. At the monthly level, calories and bodyweight showed almost no relationship — the correlation was just r = +0.05, statistically negligible. The months I ate the most simply weren't the months I weighed the most: intake stayed clustered around 3,100–3,250 kcal/day while my weight drifted across an 8 lb range for reasons that had little to do with calories.

This doesn't mean calories don't matter — they clearly do. It means that within a narrow calorie range, and with high, variable daily activity, calories alone couldn't explain month-to-month changes in weight. Bodyweight during recomposition is shaped by many things at once — activity, training, hydration, glycogen, and shifting body composition — and monthly averages smooth out much of the day-to-day signal. The honest takeaway: weight is a multivariate, adaptive process, not something a single number can predict — and high, consistent activity gave me real dietary flexibility.

Did more steps mean less fat?

Steps vs body fat: weakly positive, r = +0.45.
Monthly steps vs body fat. Correlation r = +0.45 — positive, the opposite of the naive expectation.

Surprisingly, the opposite of what you'd expect. There was a moderate positive correlation (r = +0.45): the months with the most steps tended to have slightly higher body fat, not lower. The fitted line genuinely tilts the "wrong" way.

This is not evidence that walking adds fat — correlation isn't causation. It's a textbook case of step count being a poor stand-in for energy balance on its own. Several things plausibly explain it: my highest-step months (summer) were also my highest-calorie months; more walking can quietly raise appetite and reduce other movement, blunting the deficit you'd expect; and stretches of higher body fat may simply have prompted me to move more. Body fat is shaped by intake, training, protein, sleep, hydration, and hormones all at once — so a single behavior, read in isolation on one person's monthly data, will mislead far more often than it informs.

What actually held up

1 · Consistency beat optimization

Observation: no single input correlated with the outcome, yet the outcome still improved.
Data: protein held near 169 g and calories near maintenance, month after month, for 16 months.
Interpretation: the recomp was driven by sustained habits, not any precisely tuned variable.
Implication: show up consistently for long enough and the small edges compound.

2 · The scale was the least useful metric

Observation: weight changed −0.8 lb while composition clearly shifted.
Data: body fat −1.2 pts, muscle +1.1 pts, at a near-constant weight.
Interpretation: judged on weight alone, the experiment looks like a failure; it wasn't.
Implication: track composition and strength, not just the scale.

3 · Strength was the clearest signal of all

Observation: the least ambiguous improvement was in the gym.
Data: squat 60 → 110 and bench up to 100 (plates), at stable bodyweight.
Interpretation: stronger at the same weight is direct evidence of recomposition.
Implication: progressive overload is both the driver and the most trustworthy gauge.
Squat Bench press
Squat: 60, 105, 110, 105. Bench: 90, 90, 100.
Working weight in plates (excludes the bar) — the clearest evidence the training worked.
Takeaway: the trustworthy findings were the boring ones — consistency, composition over scale, and rising strength.

What I'd tell myself at the start

Consistency beats optimization. The habits I held for 16 months mattered more than any variable I fine-tuned.

High activity buys flexibility. A large daily step base made calories almost irrelevant to weight — room to eat and still stay lean.

Recomposition is slow but measurable. The change was small and real — visible only because I measured the right things over a long enough window.

Single-subject data needs humility. Correlations from one body mislead; trends over time tell the truth.

Takeaway: the wins came from patience and good measurement, not from a perfect formula.

The repeatable recipe — as I actually ran it

15,000+steps per day (I averaged 18,699)
4–5×resistance sessions per week, progressive overload
~1.0 g/lbprotein per pound of bodyweight (~169 g/day)
Maintenancecalories — not a surplus (weight held flat)
~8 hrssleep nightly (self-reported, not logged)

Each variable earns its place. Steps create a large, flexible energy budget. Resistance training is the signal to build and keep muscle. Protein protects that muscle. Maintenance calories give the slow trade time to happen without adding fat. Sleep underwrites recovery.

Honest note: the common version of this framework prescribes a slight calorie surplus. I ran it at maintenance, and the data shows that worked for a lean, already-trained lifter — your mileage, and your starting point, may differ.

The whole picture

Weight (lb)
Body fat (%)
Muscle (%)
Avg daily steps
Final weight
163.8
Final body fat
13.1%
Final muscle
82.4%
Squat gain
+50 lb
Takeaway: leaner, more muscular, and stronger at the same bodyweight — a quiet recomposition that the scale alone would have missed entirely.