That doesn't counter his point at all. The difference between "obesity" and "maintaining weight" is still implausibly thin; what Gary was talking about in the first place was the delta in the top two lines, not the net total. He mentions that more obese people eventually do reach an energy balance, which obviously happens because they income and outflow eventually balance, which is a sign that he understands that. (Sixth paragraph, starts with "First, obese people tend to be weight stable for long periods of their life".) You didn't disprove his argument, you just reiterated bits of it, only with less comprehension and a math simulation disconnected from what actually happens (your simulation lacks the relative stability that people reach and simulates the clearly false scenario in which people become fatter without bound). ("relative stability" - there is a long-term weight gain over pretty much everybody's life but when people become obese they can do so much more rapidly than ~1lb/year)
(Personally I'd break it down to the individual meals since that's where satiety occurs and now we're talking about a delta of 3 or 4 calories. It's an absurd argument. Nothing is that precise without a feedback loop. And if a feedback loop is involved, the idea that the feedback loop is what is broken is hardly that odd an idea.)
I'm seeing a lot more people declare that Taubes is wrong or obviously a crank that there are people actually demonstrating a comprehension of his points, which I tend to consider a prerequisite for an actually-effective "debunking". Maybe he's wrong, but it's not because of people arguing against points he's not even making while blatently missing the ones he is.
Again, this is not a scientific phenomenon, it's a social phenomenon. If this was about science it would be over and done. You read his article and find it correct, I read his article and find it scientifically false. Taubes gets paid to write whatever he thinks people will latch on to. Unfortunately I don't get paid to overcome what he's obviously found to be powerful social hooks in people.
Taubes didn't sit in his study reading a "century's worth of literature" over the last decade and somehow unravel all the conventional wisdom of science with his theorizing. My personal theory, after reading his work, is that he spent much of that decade coming up with ideas to sell books. He hasn't made any advances in science. He's another Tim Ferriss/Seth Godin/Malcolm Gladwell, and if this case is anything like the past few, he'll enjoy some intellectual success and then two years later many of the slowpokes will recognize that what he's selling is enticing but not really correct.
I'm not sure what the social hooks are. Someone else suggested that smart fat people are looking for alternative explanations like in any other diet fad. I don't know if that's true, but that would explain a lot. Certainly some people are defending him purely on intellectual grounds as another intellectual defying authority, they want him to succeed. I'm glad you enjoy reading him, but, basically, he doesn't know what he's talking about.
So what's your excuse for this guy: http://www.youtube.com/watch?v=dBnniua6-oM ? PhD in the relevant field, plus actual field experience. Or the not insignificant number of other people with actual PhDs coming to very similar conclusions? (They may seem superficially different but if you take the time to understand them both Dr. Lustig and Taubes basically have the same points.)
What you appear to be aggressively declining to understand, despite two prior commenters clearly explaining it--including the one that you're responding to here--is that Taubes does not argue that 20 extra calories a day makes a person fat. He argues the precise opposite, in fact, at great length and in some detail.
What you appear to be aggressively declining to understand is that I'm arguing that Taube's straw man is misleading. I'm not claiming his main thesis is wrong, I couldn't even figure out from the blog post what that thesis is.
He used a misleading straw man: "Look, modern medical science says only 20 cals a day makes you obese! I'll repeat 20 cals/day many times, and tell you it's just a couple of gulps of beer. I'll never mention the other 100-350 cals/day (depending on year) once in 3000 words. Only 20 cals/day must be wrong, go read my book to learn the truth! Now for a made up dialogue between myself and a personified straw man!"
Yes, technically he made no untrue claims. He just deliberately failed to mention that the straw man has a good point: 7.5 sticks of bacon/day will make you fat while 7.0 sticks of bacon will only keep you fat once you are already fat.
Ahh, if only I were at a real keyboard. I'll just say that 1) you make a fantastic point in the discussion, 2)there is evidence that even when the fat person fails to eat that additional bacon, metabolism and activity levels drop to compensate, and 3) you have failed to refute Taubes' primary point, which is that "we get fat because we eat more calories than we expend" does not satisfactorily explain why we maintain a calorie surplus in the first place.
Taubes' primary point, which is that "we get fat because we eat more calories than we expend" does not satisfactorily explain why we maintain a calorie surplus in the first place.
Is that really his point? That's true, but tautological.
Similarly, the theory that "the apple I dropped falls because of gravity" also doesn't explain why I dropped the apple. All it explains is how (and not why) the apple falls.
Exactly. Every search for truth must have some aim. In the case of the apple, saying "everything falls down" doesn't get at what newton was really interested in, which was the laws of motion in general. In this case, it is silly to stop at the second law of thermodynamics when you're really interested in homeostasis in the human body.
...saying "everything falls down" doesn't get at what newton was really interested in,...
I know it's tangential, but this is precisely what Newton was interested in. The entire point of Newtonian gravity is that everything (including the moon and other celestial bodies) falls down in exactly the same way.
(Before this, you had one set of rules for celestial bodies, another for apples.)
My apologies for not being precise enough. "Everything falls down" fails as an explanation for Newton's purposes because "down" and "falls" as we use those terms on earth don't apply to stars. As a description of a phenomena, though, it points right at what Newton was interested in, as you say.
Something similar would apply here to the statement, "People are getting fat because they are overeating." As an explanation, it falls short for someone who wants to know why they overeat, but their neighbor does not.
That's my point. You are not simulating Taubes' point. You are simulating something strictly simpler and completely different. He makes an argument based around a feedback mechanism and your model completely lacks a feedback mechanism. (You and quite a few others, really.) You (all) have skipped the step where you engage with the subject and actually listen to what is being said before you cut straight to trying to disprove it.
How do you expect to convince me or anybody else of anything with that approach? I am happy to listen to arguments against his point, but you're the one building the strawmen here.
This is a brilliant graph, but I don't think you are reading Taubes' article correctly. He is not offering "straightforward calculations", rather he is setting up a "straw man" ("Aunt Sally") which he feels represents the "almost incomprehensibly naïve and wrong-headed" conventional wisdom. Like you, Taubes does not believe that 20 calories a day is a good way to represent the difference between an obese and non-obese individual.
While it's obvious from the response here on HN that he fails to make this clear, he's setting up a contrafactual argument. He's lamenting that even the best science writers we have (like Jonah Lehrer) seem to offer only simplistic truisms like the 20 calories myth. You point out (correctly) that 20 calories is misleading by using a brilliant graph. Taubes attempts (and apparently fails) to bring it down with logic and rhetoric.
Now, it's true that Taubes wouldn't fully agree with your graph. He believes that 'calories' themselves are a faulty way of viewing the world, and that the composition of the food matters more than the calculated calories. He believes it matters whether that 450 additional calories are corn syrup or lean steak. But at base, you are not disagreeing with him, and your (wonderful and to the point) graph is not contradicting anything he's saying.
I've heard this one before. "Taubes's argument is more complex than that. You need to spend more hours studying those pages, and read books x, y, and z." It's the same argument they give in college English classes for deconstructionism and all that other nonsense.
Taubes is a journalist. The people on HN are several standard deviations ahead of his audience. If he's doing his job shouldn't we all be able to understand him without any trouble? Here's a simpler argument: Taubes is wrong, and that's where the problem is.
Well, sort of. He also has an undergrad degree in Physics from Harvard and masters in Engineering from Stanford. Which doesn't mean he's right, but it does make him better qualified as a science journalist than most.
The people on HN are several standard deviations ahead of his audience.
Yes, and this is what's making me want to cry. This is one of the few audiences that knows what 'reductio ad absurdum' means, and yet bright people (whom I respect) seem unable to grasp what to me seems to me a fairly obvious argument. Are contrafactuals truly so rare as to be unrecognizable?
Ignoring the fact that his science is wrong.
Sure, it might be, but the commentary so far has not been about "his science" but about his portrayal of conventional wisdom, which he also agrees is wrong. It's like pointing out that a mathematician is obviously a fool for saying "Let us assume momentarily that the square root of two is a rational number" as the preamble to a proof.
It feels like a scene from Idiocracy. It's not that Taubes is wrong, it's that you are misreading his argument. I'm not arguing that "it's complex" --- although I'll happily stand up to defend deconstructionism --- rather I'm trying to tell you that you are in agreement with Taubes that the conventional wisdom is rubbish.
Now, if you want to point out the flaws in his actual argument and say that there is no difference between calories from carbs and from proteins, please do! But insisting that he's wrong because it really takes more than 20 calories sounds an awful lot like chanting "Brawndo's got electrolytes".
It's a new year. I want to regain my faith in the world. Please stop making me cry.
This is one of the few audiences that knows what 'reductio ad absurdum' means, and yet bright people (whom I respect) seem unable to grasp what to me seems to me a fairly obvious argument.
The problem is that his reducto ad absurdum isn't that absurd. It only sounds that way because he spends 3000 words ignoring 90-95% of the actual argument (specifically, 171 calories out of 191 in year 10).
The full argument: 7.5 strips of bacon/day will make you fat (technically fewer strips in earlier years, but you get the picture). 7 strips will maintain your current level of obesity, the extra 0.5 strips will allow you to gain an extra 2 lbs.
The argument as present by Taubes: only 0.5 strips of bacon in excess of maintenance is enough to explain the difference between an obese person? That's ridiculous!
Both of these are fair questions: is Taubes accurately portraying the conventional wisdom, and what difference in diet corresponds to what final weight maintained?
Yes, Taubes is presenting a very weak version of the argument. But I think he does this for good reason: it's the camel's nose in the tent, it's the NRA's (legitimate?) fear about reasonable gun control laws. Once he can show that the simple version can be questioned, he opens room for further questioning. And I don't think that his portrayal of conventional wisdom is that far off: we have entire industries based on shaving off a few calories from our desserts, and many people seem to believe that if the fat slobs would just park at the far end of the parking lot and walk another 100 feet they wouldn't be fat.
He doesn't present the version you offer because he doesn't believe it. You're presuming that its possible to eat an additional half a slice of bacon a day and have no other effects other than maintaining a stable weight of 2 lbs greater after a year. You're also presuming that this effect would be achieved by eating an additional sugar cube. Traditional calorie counting would also imply that you'd be in the same place if instead of the bacon or sugar cube every morning that you ate an entire 7200 calories pecan pie once a year at Thanksgiving.
Taubes says this is also wrong: bacon and sugar are metabolized differently, and will have different effects on your body. Both will also have side effects on appetite and energy level. How sure are you that they are the same? How sure are you that timing doesn't make a difference? How would you design a solution to obesity that is more effective than merely encouraging people to eat a little bit less and to exercise a little bit more?
I think these are good questions, and still open to discussion. Likely there are better approaches than Taubes', but I think he's a bright guy on the right path. At the least, I think it's worth attacking his actual thesis rather than complaining that his straw man doesn't fight back as strongly as it should.
Once he can show that the simple version can be questioned, he opens room for further questioning.
If you want to call the simple version of a theory into question, present the simple version of the theory. Don't obscure 90% of it. Taubes doesn't present the version I described because the version I described is pretty hard to question. Has Taubes shown us that we need to question the theory that 7.5 extra strips of bacon each day will make you fat?
Also, no one believes that traditional calorie counting is perfect. It's a good first approximation, but there are corrections to be made. You've identified one - you can't absorb 7200 calories in a single day. But then again, I never heard anyone in the nutritional mainstream claim that Thanksgiving dinner is what is making America fat. They just make claims like, for example, 7.5 strips of bacon/day might make you fat, and 7.0 strips of bacon will keep you fat but not make you fatter. (Substitute cookies for bacon if you want.) There are others, many of which are already well known. Pick up any sports medicine book to learn all about them.
As for "designing a solution to obesity", we already have the solution for the vast majority of people: eat less and exercise more. Taubes is trying to sell books by misleading people about the theory underlying that solution (i.e., ignoring 90% of it). It's hard to see how that is helping us get to a better solution than what we have now.
And I don't think that his portrayal of conventional wisdom is that far off: we have entire industries based on shaving off a few calories from our desserts, and many people seem to believe that if the fat slobs would just park at the far end of the parking lot and walk another 100 feet they wouldn't be fat.
You seem to be confusing mainstream nutrition science with people hawking their own wares. Modern medical science, in the form of an about.com calculator, says that you burn about 2 calories by walking 100 feet.
But then again, I never heard anyone in the nutritional mainstream claim that Thanksgiving dinner is what is making America fat.
Well, it is the opening sentence of the Jonah Lehrer piece which Taubes is critiquing:
"Thanksgiving has always been a day of delicious gluttony. According to the American Council on Exercise, the average adult consumes nearly 4,500 calories at the Thanksgiving table, which is about twice the recommended daily intake."
This, following immediately after the title "The Real Culprit in Overeating" could cause the impression that Lehrer (one of the best science journalists out there) is claiming that Thanksgiving dinner is what is making America fat.
You seem to be confusing mainstream nutrition science with people hawking their own wares.
I don't think so, although I will request a certain artistic license. I was thinking primarily of the signs up in my local Kaiser health facility encouraging people to take the stairs (rather than the elevator) in the parking garage. Mind you, I do take the stairs, and think it's a good thing. But like Taubes (who looks quite fit and trim in his publicity photos), I'm not sure this is an adequate solution to a national health crisis. Or at the least, I'm concerned that this solution is not currently working that well.
Overall, I think you're selling Taubes a bit short. If he's a hack flogging his books for sake of nothing but sales, he's at least got a better basis than most. Time will tell, of course, but I'm betting that in the end he wins the calorie counting argument, and the idea of counting all calories as equal regardless of source won't last.
From Lehrer's article: "Why do we overeat? This isn't just a problem on Thanksgiving, of course: More than a third of Americans are obese, which means that we eat way too much every day."
Regardless, Taubes wasn't pointing out that calorie counting fails on a once a year gorging incident, he was discussing a daily increase in intake. Calorie counting works pretty well for such things.
Time will tell, of course, but I'm betting that in the end he wins the calorie counting argument, and the idea of counting all calories as equal regardless of source won't last.
If that's all he is arguing, then he is doing nothing but pushing mainstream nutrition science. But if that were what he is doing, why mislead people about what calorie counting actually claims? Why not show the actual claims of calorie counting, and then show how they can be 10-20% off in some cases?
By the way, every single sports medicine textbook already does this.
What I find odd about the level of debate on this topic is that Taubes' thesis is easily tested empirically. Give up grains & carbs for a couple of months and see what happens. If, like me, you lose 30 pounds and lower your cholesterol significantly, then all the better. But if not, at least you'll be arguing from something other than "this violates the conventional wisdom." (Remember when the conventional wisdom promoted margarine and said eggs were bad?)
What do you think you have shown? Did you make sure to maintain the same caloric intake when you gave up carbs? You lost 30lbs so how do you know that your lower cholesterol was due to no carbs and not simply weight loss?
Most people on low-carb diets end up eating less calories, thus weight loss occurs (pretty basic, but it doesn't sell pop science books very well). I agree that no-carb is a viable strategy to cutting calories in a not have to calorie count way [1], but I'm willing to bet if you simply cut your calories and kept the same macro (P/C/F) percentages as before your diet you would have lost the same amount of weight.
[1] Going no-carb cuts out LOTS of the typical snack food that people eat and don't realize all the extra calories they are picking up.
I didn't go into a lot of detail in my original comment, but basically before going low carb a) I was doing a low fat, almost vegetarian diet, b) I was always hungry, c) I was losing weight, but very slowly. When I went low carb, a) I stopped counting calories entirely, b) I stopped feeling hungry (in fact, I rarely feel like snacking), c) I lost weight effortlessly (without even exercising!), and d) there were lots of other side benefits, such as elimination of the post lunch sleepies.
So, no, it wasn't a scientifically controlled experiment, and perhaps you could theoretically achieve the same results via a standard calorie restricted diet. (I'm not willing to regain 30 pounds to rerun the experiment... ;)) All I can conclude that in my case, low carb works and is much easier to maintain than a low fat (high carb) diet. And it's really easy to try.
I think you need to look at outcomes over 5-10 years, not a couple of months. As far as I can tell, radical diets pretty much always work over the short term. I'd speculate this is some combination of: you're not actually replacing all the calories you removed, and your body needing some time to adjust to the right mix and quantities of enzymes to fully metabolize your new diet.
(Also, I'm fairly skeptical that you're really losing 10 pounds of fat, or anything other than water, per month.)
Hm, this also negates what I thought was Gary Taubes's most interesting point -- why is it that obese people eventually maintain their weight rather than keep gaining. But if the maintenance need increases with body weight, there is no need to postulate any interesting regulation mechanism: just eating a constant amount of food would do it.
No, in the graph, the green line is pulling the red line up - this is exactly what Taubes says is wrong because the delta is so small. The blue line is a red herring because that's no longer the maintenance weight.
Taubes is saying that kind of precision in calorie counting is absurd, so can't be the explanation for other people that have constant weight over decades. How could a person on the blue line have eaten with such a small margin of error?
It's incredibly difficult to eat a constant amount. People generally eat to satiety, and that changes with weight. The hypothesis Taubes puts forward in the book is that increased weight causes one to eat more, not the other way around (which is conventional wisdom). So the question is, what caused the weight gain (fat storage)? Taubes argues in the book, it's carbs and insulin, not dietary fat.
I don't know why Taubes gets so much play on HN. His science is bad and his studies are cherry picked [1] for whatever narrative he wants to sell. For a large group of people who usually are critical thinkers I'm surprised at what passes for real nutrition science on HN.
IMHO, if you want to read good diet information from someone who not only knows what he is talking about, but also uses it in practice every day look up Alan Aragon [2]. I particularly like him because he is a research fanatic and understands the difference between high and low quality studies.
BTW, I am in no way affiliated with him at all, but have read his stuff for years.
[1] http://weightology.net/?p=265 -great read with LOTS of links to studies that Taubes conveniently (purposely?) ignored
Rearranging the Harris-Benedict Equation the author uses, and holding height, exercise and calorie intake constant, you get an expected increase in weight of a bit over a pound a year -- so just "eating like you're 30" alone would explain 20 of the 40 extra pounds the 50 years old in the original article is carrying...
I don't understand how author reasons that in his graph the "calories/day" for fat guy will be increasing with age, and for "lean" guy they will be decreasing.
OK, so if metabolic rate is slowing with age, I would guess the "fat" guy's graph would not be so steep, or actually would be a level line (eating the same amounts of food), while for "lean" guy the line will be somewhat downward sloping (eating less the older he gets).
(Personally I'd break it down to the individual meals since that's where satiety occurs and now we're talking about a delta of 3 or 4 calories. It's an absurd argument. Nothing is that precise without a feedback loop. And if a feedback loop is involved, the idea that the feedback loop is what is broken is hardly that odd an idea.)
I'm seeing a lot more people declare that Taubes is wrong or obviously a crank that there are people actually demonstrating a comprehension of his points, which I tend to consider a prerequisite for an actually-effective "debunking". Maybe he's wrong, but it's not because of people arguing against points he's not even making while blatently missing the ones he is.