For those of you who aren't quite so nerdy, what Devlin says in this article is that the BMI is, in its essence, a quick and dirty way of estimating the measurement of people's waistlines given that historical medical records don't usually track that information. (I've tried this out. I've divided my BMI by my waist measurement for a number of different moments in time, and I always get close to the same number.)
Devlin's article has thoroughly convinced me of the value and limits of BMI: It's like a waistline measurement, but not as good, and subject to all the limits of waistline measurement. The thing to remember with BMI is "quick and dirty" or, to quote the BBC, "cheap and cheerful." I wouldn't even dream of using it as a the
way of determining if I were healthy or fat or anything else, since I have a whole bunch of more precise measurements at my disposal for that.
Rant over. Thanks for listening.