Calibrating energy and body composition
In this investigation
Three questions that should stay separate
I questioned the assumption that I needed more than 3,000 calories while weighing around 170 lb, working at a desk and training. That was a reasonable question. A body weight and a description of workouts are not enough to establish a precise daily expenditure.
The latest food calculation estimates about 3,090 kcal, with modeled portion/composition scenarios of approximately 2,815–3,500 kcal. That interval is not a statistical confidence interval. It reflects uncertainties such as edible chicken weight and sauce composition; unrecorded food or an incorrect assumed quantity would create additional uncertainty. These numbers come from the saved food report, not a measurement of how much energy my body used.
The separate questions are what I ate, how my body is changing, and whether I am recovering well. A satisfying meal answers part of the third question and makes a routine workable. It does not directly measure the first two.
Resting metabolism is not total expenditure
A commonly used resting-energy equation was developed from 498 adults whose expenditure was measured by indirect calorimetry. It uses weight, height, age and sex; even then it predicts rather than measures an individual’s resting metabolism. Its original participants were not a dedicated sample of heavily training young athletes. Mifflin et al., 1990.
Total daily expenditure also includes digestion, movement outside exercise and exercise itself. Two people at the same weight can differ in these components. An hour inside a gym is not an hour of continuous hard work: sets, warm-ups, rest and conversation all occupy time. Scooter travel is different from the same journey on foot.
An equation can provide a starting hypothesis if its inputs are known. Applying an activity multiplier does not turn that hypothesis into a measured requirement, and the multiplier can overlap with exercise calories added separately. I should not infer either “definitely 3,000” or “definitely much less” from the current description alone.
A watch reading is another estimate
An older laboratory evaluation tested seven wrist devices in 60 people against indirect calorimetry. Heart-rate accuracy was often reasonable, while none achieved energy-expenditure error below 20%. Those were 2017-era devices, so this is not a test of current models or of a device I own. It illustrates why a believable pulse display does not automatically validate a calorie display. Shcherbina et al., 2017.
Steps, duration and training records can still describe activity usefully. If I use a calorie estimate from a device, it should remain labeled as that device’s estimate rather than silently becoming the amount I must eat back.
Weight changes are not pure fat changes
Daily scale mass includes water, glycogen, food in the digestive tract and tissue. Restarting training, changing carbohydrate intake, eating a salty meal or sweating can shift it without an equivalent change in fat.
Even a body-composition scan can be affected. In 18 trained male cyclists, short-term glycogen and creatine manipulations changed DXA lean-mass estimates during a period when substantial muscle-protein growth was unlikely. Glycogen loading produced roughly 2–3% changes in estimated lean mass, alongside changes in body water. This is a reminder to standardize measurement conditions, not a personal correction factor. Bone et al., 2017.
A quick decrease after sauna is especially unsuitable as evidence of fat loss. Likewise, a quick increase after changing carbohydrates or starting creatine is not automatically new muscle. Those changes should be annotated rather than used to overwrite the calorie plan immediately.
A practical calibration routine
I have not supplied a measured weight trend or chosen a precise body-weight target in this record. A useful first step is to state the goal: maintain roughly stable weight while returning to training, gain gradually, or pursue another explicit outcome. The following is a proposed routine, not evidence already collected.
For two to four weeks, if feasible:
- Record actual foods and amounts. Keep bought, eaten and discarded food distinct. Preserve unknown amounts instead of forcing a precise total.
- Take morning scale readings under comparable conditions—after using the bathroom, before food or drink, with similar clothing and the same scale. If daily readings are impractical, keep a consistent schedule.
- Compare weekly averages or another consistent trend summary, rather than reacting to one day’s number. Note unusual salt/carbohydrate intake, illness, travel, sauna or a new supplement.
- Record training performed, relevant duration and effort, and a short observation of appetite, sleep and recovery. A planned workout is not a completed one.
- Change one useful feature at a time when the repeated pattern supports it: an easier meal, a modest portion change or a different meal time. Then observe again.
The two-to-four-week window is a practical starting point, not a validated universal threshold. A noisy period may need longer observation. If weight is roughly stable and food records are representative, average intake provides a working estimate of maintenance under those conditions, with recording error still present. It does not establish the same expenditure after training or body weight changes.
Why simple weight-to-calorie arithmetic can mislead
Validated dynamic body-weight models account for changing expenditure as weight changes. They do not assume that a fixed calorie deficit produces the same tissue loss forever. That is why extrapolating a brief scale trend with a fixed “calories per pound” rule gives false precision. The models also depend on assumptions about intake and adherence; a sophisticated model cannot repair an incomplete food record. Hall et al., 2011, NIDDK model description.
For now, the largest gains in accuracy may come from ordinary details: verifying the yogurt label, counting nuts consistently, distinguishing a bone-in half chicken from its edible meat, and knowing how much sauce I actually used. More decimal places cannot compensate for those missing details.
Stable weight does not certify recovery
Energy balance and energy availability answer different questions. A scale trend helps calibrate maintenance; it does not, by itself, exclude health or performance problems related to insufficient available energy. Persistent unexplained fatigue, recurrent bone injuries or other concerning changes need assessment in context rather than a calorie-table diagnosis. REDs screening uses multiple observations and clinical judgment; a single day’s appetite, an estimated intake or a universal numerical cutoff is insufficient. IOC assessment rationale.
I would use the log to make uncertainty visible and decisions easier. The aim is an enjoyable routine that supports the chosen training goal, with enough repeated evidence to adjust it. Feeling slightly hungry on one earlier evening is one observation; it is not proof that the current diet or its estimated calories are exactly right.