In my trade we commonly mention the truism that “if all you have is a hammer, everything looks like a nail”. I bring this up because I’m about to do what every smart person likes to do: drastically simplify a complex phenomenon.
Specifically: If all you have is systems thinking … everything looks like a system.
Today’s subject is: how do people get fat?
A long, winding discussion of an analogy which may eventually be relevant, in which Control Systems are introduced
One of the great tributaries of systems modelling is control theory, which grew out of the study of steam engine governors.
Suppose you have a steam train. You want it to move at a more or less constant speed. To do so, you need to ensure that the right amount of steam is admitted into the pistons on each stroke.
In comes this elegant act of engineering genius:
What you see here is one of the innovations that made the industrial revolution possible. It’s a “Centrifugal Governor”, and it means that the right amount of steam will get into the pistons on each stroke. So important is this particular invention that it, along with some other clever innovations, is why the unit of power in physics is known as the Watt.
The shaft in the middle is connected to the drive axle of the steam engine through some gearing or chains. As it spins, the brass balls spin around the shaft (hardy har har). As they spin higher and higher, they compress a little pair of arms which, through a lever, adjust a valve on the piston’s steam inlet.
Courtesy of the Wikimedia Commons, the whole mechanism looks a bit like this:
When more steam is let into the pistons, the engine speeds up. The shaft spins faster, the balls rise higher, pushing the lever, which pushes the valve further closed.
The steam rate coming into the engine lowers, the engine slows, the balls spin at a lower height and the valve opens a bit.
After a while, this reaches an approximate equilibrium. By adjusting the gearing, the weight of the brass, the length of the arms, the length of the lever and so forth, it’s possible to use this simple mechanism to select how much speed you want from a steam engine.
This forms a feedback loop:
More specifically, a “balancing loop”. A balancing loop pulls the overall system towards some equilibrium when it strays too far in one direction.
Now it’s obvious that I will be talking about control systems a bit more generally in just a second, but before we leave steam trains behind, I have one observation I want you to keep in mind.
The governor ignores the internal state and configuration of the steam engine.
The governor works regardless of how the engine is currently configured. It need not account for the amount of coal in the firebox, the pressure of the steam, the friction of the pistons, the grade of the track, the amount of water on hand, the kind of coal being burnt … these are largely irrelevant. They all disappear because ultimately, all the governor needs to do its job is to observe one output of the system: the speed of the drive axle.
Courtesy again of the Wikimedia Commons, here is a classic example of a control system feedback loop:
In our train analogy, the Controller was the steam valve on the pistons; the System was the steam engine and the Sensor is the centrifugal device.
This is about the simplest possible system. Happily, it’s also the same shape as the centrifugal governor: a loop with three parts. Simple loops like these can be composed into ever more elaborate systems. Every part of the diagram can be replaced with subsystems, and subsystems of subsystems, until people start staring at bowls of spaghetti and wondering where the sauce is hidden.
The dietary control system
The simplest possible model of weight loss and weight gain is this: weight is affected by calories in minus calories out (strictly I should be talking about joules in versus joules out, but people are used to talking about calories).
Notice that we have two of the elements needed for a control system: we have a System (our body) which can be measured with a Sensor (a scale or the Mark I eyeball). What’s missing is a Controller.
The Controller is us: it’s our own selves. You look at the scale weight or yourself in a mirror, decide it’s no good, and can then change the inputs of the system (what you eat, what you do).
I am not the first to think of this analogy. Probably the best expositor of this way of thinking about the body is John Walker. In his book The Hacker Diet, Walker spends the first section discussing body weight as a control system. In fact, he does such a thorough, readable job of it that I recommend this book to anyone who wants a serious look at the bottom line of weight loss.
But … insulin! Starvation mode! Carbs! Zones! That study I saw written up in the New York Times!
Various authors have come forward in the past few years to say that “calories in minus calories out” is wrong. More subtly, they say it’s too simple. A number of additional mechanisms are then discussed that affect this or that element of body weight. For example:
- Fat storage and release
- Energy system utilisation
- Hormonal responses
And on and on it goes. Most problematically, these tend to become all-or-nothing Deep Truth About Weight Control.
In a particularly widely-read example, Gary Taubes developed a hypothesis about the relationship between dietary carbohydrates, insulin releases, satiety and fat storage that he thinks accounts for the “real” cause of obesity. Unfortunately it doesn’t quite stand scrutiny from sciency scientists.
I want you to notice something interesting: that while the internal mechanisms of fat storage and release are enormously complex, with hundreds of interacting variables, the inputs and outputs are simple.
You control what you eat and what you do.
A better diagram for bodyweight control will resemble a great big mess:
All of the hundreds of factors involved in body weight control have an impact. Sometimes quite a significant impact. But not one of these factors can overpower the simple fact that a body can’t burn dietary calories it didn’t receive.
Net calories is, therefore, a forcing element in this system. It is not the only element, but by itself it can force the overall system into a different configuration. And it’s not as though this is hard to prove. Controlled starvation studies have shown it; history has shown it (look at photos of Changi POWs) and — though it distresses me to draw a connection between acute human misery and chronic diseases of abundance — current events in Niger show it.
So final is the forcing power of caloric balance in this system, that we can potentially ignore the entire mess of subsystems in weight control and treat the whole thing as a black box. That is, we can reduce it to the first feedback system diagram above: inputs, controls, system, outputs, sensors. Calories in, calories out, weight and appearance.
Sounds familiar, doesn’t it? Just like the centrifugal governor, caloric balance requires no consideration of the underlying system apart from inputs and outputs. Calories in minus calories out, ultimately, settles the body weight balance.
Psychology and evolution
Obesity is, as I said above, a disease of abundance. It’s a disease of civilisation and particularly of capitalist civilisation, being as it is so spectacularly capable of producing calories far in excess of those required to sustain every human alive.
As Mssrs Penn and Teller observed, that we are alive today is evidence that our ancestors liked two things: “eating and fucking”. Those primeval life forms which did not avidly eat and bonk were replaced by those who did both whenever the opportunity arose. The entire mechanism of fat storage is a quantum leap forward in survivability for higher organisms, presenting as it does the ability to average caloric inputs over time and to decouple the organism from short-term fluctuations in food availability. Without fat storage animals could not have evolved to their current level of complexity, because that complexity imposes a very high cost in calories. And humans are no exception.
So it’s understandable that people don’t want to control their eating. Or rather, it’s understandable that the abstract, long-term promise of weight loss is swamped by a hindbrain screaming “EAT THE CAAAAAAAAKE, ALL OF CAAAAAAAAKE”. Much in the same way that cars pit a visual-motive system adapted to running speed in the African savannah against tons of steel moving at dozens of kilometres per hour between buildings only metres apart, weight control pits us against our own adapted nature.
Nevertheless, we are not automatons. We have higher brain centres, which have evolved to the point where they can exercise abstract reasoning and suppress our instincts. Diet is no exception to this happenstance. Unless you have diminished psychological functioning (been there, done that, got the pills), you are in theory capable of taking this step. You are not a victim: you are a master of your own destiny.
The key to weight control is finding the dietary template that you can stick to for the rest of your life. That will be different for different people. Some folks are happy with calorie counting a la John Walker. Some people are happy with paleo, primal or ketogenic diets. Others swear by zones or low fat and so on.
Worrying about why a diet works for you is irrelevant. That it works and that you can sustain it indefinitely is what matters. If you can sustain a caloric deficit for the time it takes to drop to a healthy weight and then stabilise, then that’s the diet for you. Crash dieting or yo-yo dieting is not OK.
What about people with thyroid disorders? What about this low metabolism thing? What about … and … also … but!
Every time you point out the simplicity of the net calories model, somebody comes up with special pleading.
I distinguish special pleading from non-caloric causes of obesity because they tend not to be theories about how obesity occurs. Rather, they’re virulent memes which allow us to turn that powerful, higher abstract reasoning onto the path of justifying what our hindbrain wants us to do.
My objection is that very few people have a genuine reason why simple calorie counting won’t work. Very few people. Yet everybody is quick to self-diagnose. “Oh!” they say when they hear the latest health journo blurb. “I must have that condition. I guess I’ll never be able to lose weight. I should just accept it. Pass the gravy, my glass is empty.”
Well look. Maybe you do have Schnitzengrüber Syndrome. But do you know what? You probably don’t. And until you see your doctor and check, you can probably assume that you don’t and that putting the fork down is the bottom line.
And … even if you do have a medical condition which affects your particular biochemical reaction to food or exercise, it doesn’t change that net calories is still a forcing element. People with every metabolic malady known to science, and a few more that aren’t, will have died throughout history through simple starvation. No disease can create energy from nothing.
One thing I liked about the Hacker Diet is that Walker teaches you how to control yourself. Make frequent measurements and then adjust your system inputs. No need to make dramatic changes, folks. Weigh yourself daily, perhaps take weekly photographs for visual comparison, and gently reduce your calorie intake.
Because you’ve kept the records of weight and appearance, you can see what your metabolic system is doing. Why speculate about how you will respond to food? Measure! Verify! You are a control system.
Caveats and Conclusion
What annoys me about models like Taubes’s is that on the one hand they complicate (“It’s not just calories in minus calories out!”) and then over-simplify (“It’s caaaaaaarrrrrbbsssss!”).
Well, if it please the court: I’m guilty of this too. I’ve simplified enormously. Abstraction is the basis of my profession (software engineering). But that doesn’t change the fact that the model I’ve outlined above is true, by any serious empirical standard. Cut calories enough and you’ll lose weight, add enough and you’ll gain. How much those inputs affect weight and appearance will vary wildly from person to person, but the underlying relationship will always be there.
If you have one, and only one thing that you take away from this screed, it’s this: calories in minus calories out is a forcing element. It will always, in the end, force body weight in one direction or the other.
But there are caveats! Simply bludgeoning the body into submission with help from that pitiless prick, Mr Physics, may be counterproductive. For example, cortisol is a hormone which inter alia mediates fat release. But it also mediates protein breakdown in muscles and the density of bones. Crash dieting tends to kick cortisol up a lot and well, it’s not too picky about what receptors it plugs into. Crash dieting isn’t healthy because as well as losing fat, you can lose muscle protein and bone density also.
Another caveat is mental health. As has been raised by people I’ve argued with on Reddit, depression can affect the will to seek help or the effectiveness of “tough love” advice. And, of course, for some people, calorie counting or crash dieting spirals off into serious health issues such as anorexia, bulimia or (less seriously) orthorexia. In such cases professional help should be sought first of all for the mental health issue. Weight can wait.
So. Here’s the moderated view of what I want you to take away:
- Calories in minus calories out is the ultimate, unavoidable arbiter of body weight.
- But don’t crash diet. Lose weight gradually over time. Pick a diet you can enjoy and sustain for the rest of your life. There’s lots to have a go at.
- Don’t lie to yourself. Keep weight and photo records to know you’re making progress. If you get thoroughly stuck and your records show you’ve been doing it right, see your GP and/or a registered dietitian (not a nutritionist, these are frequently unlicensed quacks) to see if you have something interfering with your metabolism.
- Do some exercise, and in particular do some training with weights to prevent muscle and bone weakening.
Patience, friends. Patience and consistency can achieve amazing things. Trust me.