Where these charts come from
The WHO's Multicentre Growth Reference Study followed over 8,000 healthy children from six countries (Brazil, Ghana, India, Norway, Oman and the United States) from birth, raised under conditions considered optimal — mainly breastfed. It is not an average of how children generally grow, but of how they grow when nothing limits their growth: that is why the WHO calls it a 'standard' rather than a plain 'reference'.
That has a practical consequence: a formula-fed baby can show a somewhat different weight pattern than the original cohort's without it meaning anything concerning. The chart remains the right one to use; it just helps to know where it comes from.
What a percentile actually is
The 50th percentile is the median: half of children that age and sex weigh more, half weigh less. A 15th percentile does not mean 'underweight': it means that, out of 100 healthy children that age, about 15 weigh the same or less. It is a position within normal variation, not a grade.
The exact figure is calculated with the LMS method — the same one the WHO uses to draw its own curves — which converts weight into a z-score (how many standard deviations from the median) and translates that z-score into a percentile.
What actually matters: the trend, not the point
A single percentile, on its own, says little. What a paediatrician watches is the curve: if a baby has been at the 40th percentile for months and stays there, that is going well — growing steadily at their own pace. What raises a flag is a sharp change: crossing two or more major percentile lines downward (say, from p75 to p15) in a short time, regardless of where they started.
This calculator, like any tool fed a single measurement and date, cannot see that trend: it only shows today's snapshot. If a trend is what worries you, the useful thing is not this one figure but the curve recorded across the child's check-ups.
Head circumference, a separate measure
Head circumference follows its own chart and its own meaning: it is not as tied to weight or nutrition as an indicator of brain growth. A low, stable percentile is not concerning by itself if the rest of development is on track; what does warrant a check is a sharp change in percentile, a measurement well below p3 or above p97, or a head that has not grown over several consecutive months.