Oh, where to start. - BMI is a poor measure of obesity. Among other things, bone and muscle are both significantly denser than fat. People, including childre

Oh, where to start. - BMI is a poor measure of obesity. Among other things, bone and muscle are both significantly denser than fat. People, including children, with thicker bones and/or more muscle will have higher BMI regardless of fatness. So it's at best a poor proxy for fatness, not the One True Measurement that so many people use it for. - BMI is an especially poor measure for children. The ideal BMI (to the extent that means anything) varies according to developmental stage. Which is why you'll get tables and graphs for BMI that are indexed by age---except that age and developmental stage are themselves only roughly correlated, and a child at an earlier or later stage than is typical for their age will not be well-described by such a chart, and the "ideal BMI" that the chart lists for that child's age will actually be a non-ideal BMI for that child. - BMI percentile is an even poorer measure, because it provides only a comparison against the population; if the population is skinny, even a high percentile might not be overweight, and if the population is fat, even a low percentile might be overweight. And that's modulo all the concerns in the previous two paragraphs. - Even after taking all of that into account, we still have the problem that our conventional idea of "overweight" meaning "more weight/fat than is healthy" is likely inaccurate, in that some amount of certain kinds of fat actually make one healthier, and (as cited elsewhere in this thread) people who fall into the range often called "slightly overweight" actually live longer. So evaluating kids' health (or their parents' assessment thereof) with reference to their BMI percentage is fraught in four different ways with inherent, systemic problems.