Sunday 24 June 2012

Starvation, diets etc. for no good reason

I wanted to post something that's occupied everyone's mind at one point or another: losing weight. How does it work, this whole fat burning thing? How much fat is good, how much is bad? I'm not professionally educated in this domain, but that's not important since a lot my criticisms come from simple assumptions. Recently I read an interesting statement from the e-book 'Experiments with Intermittent Fasting' by Dr. John M. Berardi:
the standard North American diet is often hyper-energetic – we eat more than we burn – which leads to weight gain over time.
A good test to see if a statement makes any sense is to declare the opposite and see what comes out. In this case I came up with "A non-standard North American diet is often hypo-energetic – we burn more than we eat – which leads to death". If I had to choose sides, the former certainly looks healthier than the latter.

Even more interesting was that Dr. Berardi goes on to claim that intermittent fasting (in his definition not eating for 20-odd hours) can reverse these 'problems':
Intermittent fasting can be helpful for in-shape people who want to really get lean without following conventional bodybuilding diets, or for anyone who needs to learn the difference between body hunger and mental hunger
Unfortunately this makes no long-term sense: if you burn more calories per day than you eat, eventually you will die. This is not science; it is simple logic. For even a slight negative imbalance your fat reserves will deplete, which could take years, but at some point the reserves run out.

There are two variables to work with: calories in and calories out. Calories 'in' is the simpler of the two variables to measure, and in theory should be very exact. But it's not that simple. Imagine, if you will, that you eat four times per day and that each meal had a 100 calorie uncertainty. Think how small 200 calories is when compared to an entire meal (some beers are more than 200 calories apiece). Perhaps you ate an extra dinner roll, forgot to count in a glass of wine, or omitted a snack. These are extremely fair assumptions especially since you have to also include the errors inherent in calorie (mis)labeling in food.

Let's take this 100-calorie imbalance to the nth degree. If the mistakes are completely random (i.e. you miscount above or below the true value), then summing the errors for the day we have

Total error/day = sqrt(sum of squared errors) = sqrt(4*100^2) = 200 calories of uncertainty

Now 200 calories isn't that much to mistake per day, but unfortunately you don't get to reset this value every time you go to sleep (why would you?). As the weeks go by, the errors increase. Imagine a full year of this uncertainty:

Annual food intake 'error' = sqrt(365*4*100^2) = 3820 calories

For an entire year 3800+ calories is also small (maybe what you'd burn in a marathon). 3800 calories is close to a pound of fat. Let's change things a little. Since it's much, much more likely you'd overeat than undereat (after all our brains are programmed this way), the true error is biased. A biased error means that each mistake is made in the same direction. Therefore it would be more prudent to add together the uncertainty of each meal in the same positive direction. So suppose you unexpectedly ate a mere 50 calories more than intended for each meal (and 4 meals/day or 3 meals and one snack), then in a year's time....

Annual food intake 'error' = 365*4*50 = 73,000 calories (!!)

A measly 50 calories of 'overeating' per meal equals over 20 pounds of weight gain in a single year. The math is simple but I bet you never thought about it that way. And this calculation applies to everyone, including active people. Athletes burn more calories but few people seem to remember athletes also eat more. In fact they eat so often (and eagerly) their intake errors are expected to be much higher. According to these calculations everyone should be fat. While writing this entry I ate a 'personal sized' apple pie that turned out to contain over 600 calories. I do that sometimes. But my weight has remained unchanged for the last five years. Something doesn't add up here...

In the case of athletes, recall that different kinds of athletes have different body shapes. This will be true even if each burned the same number of calories a day. When exercising, it is the sport itself, along with your body's internal regulations, that will determine your overall physique. Doing a lot of body-building exercises will make you look like a body builder. I lost 20 pounds of fat and muscle when I switched from XC skiing to running (and have remained the same weight ever since).

How the body does regulates all these inputs is a mystery. A really cool mystery, but imagine you were your body receiving all these signals (food, energy losses, sleep). Odds are you'd prefer to keep in equilibrium for as much of your time as possible. Homeostasis is connected to metabolism, which is variable. Not eating for very long periods makes you sluggish to save energy. Some of that sluggishness is mental but with time it is quite physical. But assuming you eat enough, what exactly you eat and precisely how much of it has far, far less to do with this overall process. You can't reverse engineer yourself by simply changing your diet. By measuring someone's eating habits can I really know that much about them? How well could you guess the make and model of a car just by knowing how much gas the driver puts in the tank? (NB: the tank might not be empty!). Long-distance swimmers have more body fat than sprint swimmers but long distance runners have less than either. Do you really think this is a diet-based problem?

Among entire populations there's a similar story. When discussing excess weight, few people seem to realize that those who are fat also have a stable weight for long periods (that is few, if any, people keep piling on weight forever). If you know someone with a bit of heft to them they probably have had the same amount of heft over a number of years. Body weights do tend to creep up over time with inactivity but not at the rate I alluded to earlier (i.e. 20+lbs/year, every year). Whether active or inactive, the body has made rough a decision where it's default weight should be.

But not everyone has the same default weight; some are thin, other not. Why? Simple: throughout human history it's been an advantage in periods of starvation to put on weight easily. But at other times it has been important be lean and mobile (i.e. for hunting, historically). To prevent total population annihilation, some regularly active people have a disposition for surviving lean times while others are prepared for active times.

A handful of thin people might survive a famine by slowing their metabolism to an extreme lethargic state. It will keep you alive, but not active. All this is just our 'default' setting; it is a weak predictor of leanness when habits change. For instance there are obesity epidemics in Northern Quebec Cree communities due to inactivity. Yet a single generation past these same communities were lean when hunting using traditional means.

Our modern sedentary history is a blip on the time scale of humanity. A completely sedentary lifestyle is unprecedented, so much that we have always assumed some degree of physical work to earn our food. This causes unexpected problems for your body the same way a rodent's teeth are expected to be worn from daily use (hence constantly grow). For beavers not to chew down trees would mean eventual death by overbite, though unlikely given their habits. For the most part humans were 'meant' to exercise simply because that's how our software/hardware was designed. Some animals 'need' more exercise than others. But... starvation was never intended as a personal 'control' parameter. To lose weight by starvation is obvious. That this weight will return with a vengeance is also obvious. For such regained weight to be again eliminated by further starvations is, you guessed it, obvious (and stupid).

Two conclusions, both not in favor of Dr. Berardi's ideas:

1. Humans do not benefit from fasting. But yes, we are certainly well equipped to survive it. Even a lean athlete has around 8% body fat, enough to run him or her hundreds of miles. Elementary math can demonstrate as much. Dr. Berardi's analogy is broken: we do no need to replicate past starvation. Instead we should imitate our past activity levels. We do not require a literal imitation of our hunter/gatherer lifestyle (we don't actually know quite how they lived on a daily basis). This much we know: some eating here and there (wherever food can be found) until full while doing spurts of cardio for several hours a day is quite normal (again, historically speaking). And yes, being a little bit hungry helps you think sharply for a few hours (surprise...). But eventually it's all you're thinking about (Again, surprise...)

2. Calorie counting does not work. Eating fewer calories than you burn means death (or a decreased metabolism and increased lethargy). Because you cannot 'exactly' measure the calories in meals (as even the tiniest errors creep up until they reach absurd proportions) the notion of over-eating is a trivial one. Unless we ate exactly the correct number of calories, which we can't, it is more appropriate to say we all over eat. Especially athletes. The body pre-chooses a weight based on exercise-related inputs and makes bodyweight adjustments by variable metabolism (that post meal warmth you might feel is real). To do no exercise at all screws up the body's equations and it's anyone's guess what will happen, but probably gain a lot of weight, but you might be thin by default. By modern standards you're lucky if your default is something lean (given the current availability of food). Those same people are not so fortunate if food ran out for a whole month.

Having recently earned a PhD myself I'm less than impressed with Dr. Berardi's credentials; I have no illusions that earning a doctorate alone means anything intelligence-wise. After looking at his book I see a man with curiosity, which is good for research, but conclusions which are questionable. Not eating will, in the very short term, mean weight loss, and in a long-term way signal to your body "hey, I'd better stock up on more fat next time". This, in itself, does not require a PhD to figure out.