These studies were very empirical, and they really were aimed entirely at determining how much to charge middle-aged white people for insurance. They did a good job of that, but they didn’t account for any number of confounding factors. However, once they had published the tables, these tables went from being essentially descriptive — “middle aged people seem to live longest in these height/weight ratios” — to be taken as prescriptive — “everyone’s ideal weight is given by these tables of height/weight ratio.” Now we define these “ideal weights” in terms of body-mass index, BMI, which sounds much more precise and scientific, but turns out to be simply a height/weight ratio.

Pretty much anyone can see that BMI is questionable — for example, a champion bodybuilder with a competition body fat of 3 percent may well have a “very obese” BMI. (On the other hand, it’s unclear that very low body fat is necessarily healthy either — in fact, we know it’s certainly not healthy for women.)

The problem is that epidemiology, the study of health and disease across large populations, keeps finding results that don’t quite fit this idea of ideal weight. Most recently, a study by Katherine Flegel and others published in January of this year showed that the notion of ideal weight was massively oversimplified. The study found two things: first, that for younger people, BMI doesn’t have any strong effects, and as you get older, the BMI associated with the least likelihood of dying increases.

In other words, if you don’t want to die the data suggests you actually want a slightly higher BMI as you get older.

What’s more, other studies say that BMI isn’t as good a predictor as simply the length of your belt — the larger your waistline, the more likely you were to have a whole lot of different health problems like type-2 diabetes. (This one does fit the bodybuilder example, too — bodybuilders do generally have small waists.)

Now, this can be taken too far — there’s no doubt that real obesity has bad effects on your health. (My knees would tell you that, if asked. And if knees could talk.) But the truth is that being a little overweight is either not harmful or may actually be helpful.

Myth #2: Your Weight Depends Only On Calories-In Versus Calories-Out.

This one has also come up here before. My own data, collected over more than six months and my first two 13-weeks experiments, has shown that it certainly doesn’t hold for me. Over the first about 10 weeks of my very low-carb diet, I lost weight much more rapidly than could be accounted for by a calorie deficit; since then as I crossed the Great Plateau, my weight loss has been considerably less than calorie deficit should predict.