Calories, however, are so irrelevant to body composition and health that they should rarely enter the equation at all.

I know this thread is a day or two old, but maybe someone will stumble across this and learn something.

First of all, measuring the caloric value of our food is a flawed concept. Until heat is released in a chemical process, the number of potential calories in a substance means very little. Wood, for instance, has plenty of "calories" because you can burn it when it's dry, and it burns quite well. Yet what does it mean to consume say 200 calories worth of wood pulp when no nutritionally relevant amount of it will result in heat being generated by the metabolism? Fine, so we don't count plant fiber as having calories on food labels, but as I explained in another comment, macronutrients like fat, protein, and carbohydrates are not equivalent even if you were to consume each of them independently with the same number of calories. Your body is going to compose itself differently on a high fat diet than it would on a high carb diet, or a diet that combines fat and carbs.

A major problem with CICO is that we cannot reliably measure the "calories" in our food, despite having those calories listed on food labels, because those calorie counts are allowed to be incorrect by as much as 20%. That much variance means that any calculation someone tries to do with calories cannot be done with confidence. If one believes they should only eat as many "calories" as their basal metabolic rate, then they cannot know whether they are actually consuming hundreds of extra "calories."

The only way someone could come even close to accurately measuring the calories in their food is if they were to take each component, dehydrate it, and test its heat output in a bomb calorimeter. Long story short, that's not going to happen.

So the idea that calories are a useful measurement for food intake is a failure. That is, unless, it's done in a laboratory setting with the correct kit being used by a team of qualified researchers, and even then it can be difficult for them to get it right.

Now let's talk about calories-out.

Though it is possible to measure the metabolic rate of a human by measuring their body heat in calories, this information is generally not actionable by most people. Measuring metabolic rate tells us virtually nothing about a person's health or how their body is composing itself; let's face it, most people care more about having less fat and more muscle than they do their metabolic health. A muscle-bound person and an obese person can have a similar BMR because metabolic rate is primarily dictated by body mass. Yet there are variances between even people of equivalent body composition, meaning that using tools like Harris-Benedict are an educated guess and nothing more. When people use formulas like the Harris-Benedict formula with the belief that it's a reasonable approximation of their BMR, there's a good chance that he result of the equation isn't accurate. Again, it's a tool with limited use and little to no application for an average person. Yet, there are all sorts of BMR calculators on the web and articles telling people to use them. They are a complete waste of time.

Fitness watches don't measure calories, either. Again, they are an approximation that can be way off because there are metrics they fundamental aren't capable of capturing. At best, they can be a proxy measurement for physical activity, but they cannot tell you about your body composition or whether your actual metabolic rate is slowing down in response to physical activity. Anyone who says otherwise likely doesn't know what they're talking about.

So how do we actually measure the calories of heat coming off a human? We measure the goddamned heat, and all of it.

Direct calorimetry can be used to measure a person's metabolic rate. It involves either placing the subject in an insulated room that's surrounded by a jacket of water, or even placing them in an insulated bath of water, and measuring the temperature change. Again, this is not providing much useful information to a patient, and it involves equipment that isn't accessible to most people.

We can attempt, however, to measure both calories and metabolic state by using something called indirect calorimetery. This involves measuring the exhaled breath of a patient, looking at the ratio of oxygen to CO2, and making some estimations on metabolic rate and the ratio of fat to carbs they are metabolizing. Though indirect calorimetry might be of some value to a patient, it's not a trivial measurement to make. It requires very special equipment that needs to be calibrated and operated by specialists. I know this because I built an indirect calorimeter; 99% of people are not qualified to use one even if they could afford the $8,000 price tag. For greater accuracy, the patient's urine would need to be measured for nitrogen. And even then, there are flaws in the approach that can only be revealed by a real-time high resolution X-ray procedure, which is cutting edge and not something that really anyone has access to or could even begin to know how to use.

As I alluded to earlier, even if one could get an accurate calorie reading for a given moment in time, that doesn't mean their body won't then try to compensate for a sudden expenditure of energy in the form of intense exercise. Many people are under the impression that earlier humans must have stayed thin because they were burning so much energy all the time hunting and foraging, but research on the most "primitive" tribes of humans we know of today doesn't support this. It's never in the best interest of an animal to be constantly expending their energy, which is why they tend to conserve energy for when they need it. Early humans couldn't afford to be constantly running or whatever because it's a waste of energy and potentially deadly. Metabolic rate can even slow down after an intense bout of exercise because the human body has an interest in throttling back energy expenditure when it's no longer required.

Calories basically just suck.

When people say that they lost weight by calorie counting, they have no idea what they're talking about other than that they know there was some correlation between the calories they were counting and the weight that they lost. Your average person doesn't keep track of their actual dietary changes they're making unconsciously, and almost nobody bothers to measure body composition. Losing fat and wasting muscle is very different from preserving or building muscle and losing body fat. Like I said, most people want to lose fat and don't want to look like a string bean. Someone trying to lose weight will be in the mindset of avoiding the jelly doughnut, and eating a salad because it's "low calorie", but what's really going on there is not the calories but the change in hormonal profile. As I mentioned elsewhere, the body does this not because it has an internal calorie meter, but in response to the kind of food you're eating and the demands of your cellular processes.

On top of that, the idea that anyone can or should measure their metabolic rate is basically hogwash. Yes, you're measuring something, but outside of lab conditions, it will neither be sufficiently accurate nor actionable in the way that you think it might be.

The best that people can do is abandon the CICO model, change their food composition and partake in brief, intense, and infrequent exercise. If they really must have a quantitative measure of their food, they can measure its mass, but that would only make sense if they're taking those other steps and the only factor left over is that they might be overeating. To clarify the overeating part, that's still not about calories. It's possible to eat too much protein such that too much of it goes towards glucose production, or you might be eating it too frequently, or perhaps you just want to take fewer shits during the day.

RIP CICO.

EDIT: I have verbal diarrhea about this topic because I once had the hubris to think that I could fix obesity by making accurate calorie-counting available to anyone.