2008-01-20

A Skeptical Mathematician Loses Weight by the Numbers


I am six feet tall. A few months ago I weighed 190 lb, which is at the 38th percentile of American males. In other words, 62% of males my height weighed more than me. Sounds good, but there's a problem. Says who? Says WHO, that's who. The World Health Organization (WHO) has this nasty thing called the Body Mass Index (BMI) and if you are six feet tall and 190 lb you have a BMI of 25.8 and are deemed to be moderately overweight. My doctor (who is skinny) thought I was moderately overweight too.

This is the story of my journey to figure out how to lose some weight.

For no good reason, I decided I should weigh 170 and it would be all right to take two years to get there. How to do it? We all "know" that diets don't work. (Actually, we don't know much of anything. See Good Calories, Bad Calories.) There is a school of thought that if you just practice intentional eating, or mindful eating, you will eat less. A quick read of Mindless Eating (and Influence) convinced me otherwise.

The more I looked into weight control, the more skeptical I became that anyone knows anything about it. The proliferation of zillions of fad diets seemed all too understandable: diet studies are hard to do with scientific rigor, no one does them right anyway, everyone has a vested interest, etc. etc. The state of the science for all this is depressingly poor. Sigh.

Everyone seems to agree that reducing calories will result in weight loss. That doesn't mean it's true, but everyone agrees. So I decided to count calories. Being a mathematician, I was getting interested now that it was becoming a numbers game. The standard wisdom is that you need a certain number of calories to maintain your weight and if you eat 500 calories a day less than that you will lose one pound in a week. Counting calories is pretty easy because nutritional info for generic food is readily available and (thanks to Ralph Nader apparently) there are calorie counts on pretty well all packaged foods.

So I just needed to know what my maintenance calories were and then eat less than that. The maintenance level is loosely known as the Basal Metabolic Rate (BMR). Fortunately there are many Web sites that tell you how to calculate your BMR. Unfortunately, they all give wildly different answers. But fortunately, I live in a quasi-socialist society where they think the government is supposed to help people and they have a Dial-a-Dietitian service. Forget all the confusion on the Internet, I could ask a professional for the answer!

The dietitian was very nice but was very insistent that she wasn't going to tell me what my BMR might be. Why? "Diets don't work." Well, I'm not really planning to diet exactly, just practice a little portion control perhaps. "You'd have to keep track of your calories and no one does that effectively." I can! I can! "You should try to get more exercise instead." I will, I promise, but maybe I could combine that with a little calorie monitoring? No luck, resistance was futile.

So much for theory. I decided to just experiment on myself, eat a little less, count the calories I was eating and see what happened. Well, a funny thing happened. Counting calories made me aware of how many stupid calories I was taking in. I ended up eating even less than I had planned to and my weight fell faster than I had expected it would. The best part was I didn't feel like I was dieting or starving myself. Somewhere along the way I decided that my target weight should be 162 lb because that would give me a BMI of 21, right smack in the middle of the healthy range. That number was kind of arbitrary, but I liked the idea of performing a "random act of self-discipline" (as Bill Gates supposedly described his temporary adoption of vegetarianism).

The graph above shows the results. Before my little calorie-counting episode, my weight was slowly creeping up. Afterward it went steadily down. I'll leave it as an exercise for the reader to determine at what point in the graph the change happened. Just last week, I reached my target weight. I had lost 28 pounds in about five months. Yay!

What did I learn? Lots of things, but they may well apply only to me and no one else. Calories do make a difference. Fat, not so much. My normal diet is not particularly low-fat. Nor is it particularly low-carb. And exercise? Surprisingly, it seemed to have very little impact. For the first third of the five months I was exercising much more than usual. In the middle third, I exercised much less. In the final third, I've been back to normal for me, which is on the light side of moderate. None of this made any difference to my weight. (Ha! Take that, professional dietitian!) Basically I learned that 99.9% of everything you read about dieting, exercise, calories, carbs, fat and everything else is garbage. Evidence-based medicine is a good thing. But you'll be disappointed if you think there is much of that going on in the world of nutrition.

I know what you're thinking. Will he be able to keep it off? I'm actually more concerned about something else: how do I stop losing weight? I know how to lose weight. I don't know how to maintain my weight. I know what the theory says about my BMR. I can put all the data I have collected about myself into a first-order inhomogeneous differential equation and solve that to get my BMR at any given weight level. (Parents: your kids wonder why they should study math? There's your answer.)

But according to that calculation I would need to eat 700 more calories per day to maintain my weight. Um, that would be like a 47% increase. I don't believe it. So I'll experiment some more and explode some other theories in the process no doubt.