Hacker News Reader

> if they are so unusual, doesn't that make confounding more plausible?

That is a very good point.

> Also very interesting, and counter to the usual narratives about health. (Everything is heritable...)

And yet this need not lead us to the conclusion that self-improvement is futile. Consider the scenario where diet is highly heritable and also extremely important to health outcomes. If people could be convinced of the latter, they would, to varying degrees, improve their diet and therefore their health. After smoking was tied to lung cancer, I imagine that people (at all levels of genetic predisposition to smoking) greatly reduced the amount that they smoked. Smoking levels could have been 100% heritable both before and after the revelation, but would have been reduced over time.

> 5. fat percentage: -3.3 (!)

I'm not sure why you singled fat percentage out as the least significant (in the colloquial sense). The -3.3 is a reduction from 24.0, so the percentage change is ~14%.

> 2kg of weight? I fluctuate more than that on a weekly basis...

To take your quip seriously, I don't think this is a good comparison. If a specific intervention increased lean mass and decreased fat mass by equal amounts (say, 10kg), it would be highly beneficial and yet count for absolutely nothing with respect to BMI or weight. The specific changes cited here are 1.4kg lean mass and -3.3kg fat mass. I would actually count that as a 4.7kg improvement!

I wouldn't count it as 4.7kg improvement. 3.3kg is fair though. That would imply bodybuilders are close to 50 kg "healthier" than I am for the same height assuming no fat loss. But I think none of these measurements are that great to prove any point. Are we optimizing for longer lifespan, attraction for the opposite sex or just some arbitrary number goals?

Maybe 50kg is too much, but you wouldn't be happy with a 1.4kg increase in lean mass? I assume that most of that is an increase in bone density and muscle mass, both of which are good things.

These people are early- to mid-20s. Americans generally get fatter steadily from their 30-60's. So, the differences are likely to compound. Also, as a country Finland is more fit, so it's likely the gap would be wider in say the US.

  Finland Adult Obesity rate: 23%
  America Adult Obesity rate: 35%

> I would actually count that as a 4.7kg improvement!

When I quoted 2kg of weight, I meant '2kg of weight', not anything more complex. That's the total change. The exercising twins weigh 2kg less, total, all in all, on average. (If you think you've found a twisty interpretation of quoted figures which is more positive than it looks, please check the fulltext first - that's why I provided it!)

Yes, I understand that 2kg is the total change, and I've read the full text that you've helpfully linked. I'm not sure what your point is?

>> Most of the pairs had maintained remarkably similar exercise routines, despite living apart.

> Besides the testament to how 'everything is heritable' inherent in that observation [...]

>> Interestingly, the twins tended to have very similar diets, whatever their workout routines, so food choices were unlikely to have contributed to health differences.

> Also very interesting, and counter to the usual narratives about health. (Everything is heritable...)

Why do you think this proves genetic causes? They were twins, after all, they were both socialised under the exact same circumstances for two decades. Given they had identical upbringing, I'm not surprised that e.g. their preferences to food are quite similar.

This is certainly the stronger assumption. Eating habits in particular are very social and habitual. It doesn't stretch the imagination to imagine the twins, having lived in a single household for >=18 years, have a deeply set familiarity with certain foods, recipes, etc.

Because shared environment effects are typically small or trivial. You don't need a formal analysis to have a high confidence that the food preferences will be partially or mostly heritable, just like everything else. In particular, we already know that anything to do with obesity is extremely heritable and shared environment matters far less than everyone believes: see the studies in https://jaymans.wordpress.com/obesity-facts/

Thanks for the response, I'm just reading through the linked meta analysis.

It's important to remember, though, that correlations measured in twin studies, by design, only say something about the difference of both measured variable expressions, not about the variable itself. So if twin A has a BMI of 30 and twin B has a BMI of 35, that linked meta analysis says that heredity explains 75% of the difference between 30 and 35, or only 3.75 BMI points/ 10% of twin B's BMI in the given example. Even more interesting: The older the twins are, the smaller the difference gets (grep for 'decreased mean age' in the study).

You (and the study) are treating exercise as if it is some homogenous substance, and the intensity and type of exercise had no effect on the outcome. To quote the study:

"the so-called active co-twin of the twin pair was physically active Q2 times per week, and the so-called inactive co-twin of the same pair, e2 times per month (inclusion criterion 1 is shown in Fig. 1). If this criterion was not met, the physically active cotwin needed to participate in leisure time physical activity Q2 times per week at an intensity equivalent to easy or brisk running"

No-one who is serious about training would expect any body composition changes from an easy run twice a week, except at the very lowest levels.

As others have said, "you can't out run your fork". If fat reduction is your goal, you have to get your diet in order. This is generally accepted within the fitness community, and it's a straightforward consequence that is trivial to consume more calories than you expend unless you're, say, trekking in the Arctic. If you want to increase your lean mass you'd better be doing resistance exercise twice a week at a minimum and more like three or four times as your training level increases.

> No-one who is serious about training would expect any body composition changes from an easy run twice a week, except at the very lowest levels.

What's wrong with that as a minimum cutoff level?

And if that's how people actually exercise, then the inefficacy is important to know. This reminds me of the arguments that 'communism has not failed, it's never been tried!' (Real exercise has not failed, it's never been tried by these twins...)

My doctor has told me one thing over and over.

Want to lose weight, eat less and eat right

Want to look good, exercise and drink water

Somewhere down the line he stated that exercise tended to help you live longer and experience less issues as you grow up but it won't help you dodge that bus so enjoy yourself and don't get fat

Good point about the "significance" thing.

However, the word "exercise" probably covers many different regimes here. If you take a pair of twins, put one on the couch for 3 years, and give me the other to train for 3 years with Starting Strength/Texas Method or similar, I will make him into an unstoppable monster in every dimension you can measure: Bone density, neuromuscular function, 1 rep max squat performance, bodyfat percentage, lean bodymass gained, whatever. He will even look conventionally "ripped" by the end of it if you let me control his diet a little bit. Same story if you subject him to the training regimen of an elite sprinter or something like that for 3 years.

If that same twin does Nautilus machine circuits for three years, he will probably be ~pretty much unchanged, depending on various genetic factors pertaining to sensitivity to light exercise.

Probably all of the options above would be colloquially called "exercise," but they're not even approximately created equal.

    > if they are so unusual, doesn't that make confounding
    > more plausible

What does confounding mean in that sentence? Anomalous results on small sample size?

Imagine it was really hard to get two twins with such different exercise patterns, unless one developed a brain tumor. A study showing the twin with less exercise was less healthy wouldn't mean much.

A confounding variable is a variable that affects the outcome of the test but is unknown to/ unobserved by the researchers.

Example: You measure the effect of alcoholic beverages on general health. You create two groups, one group consists of the beer drinkers, another group gets the wine drinkers. Result: drinking wine correlates with better health, drinking beer does not. The confounding variable in this case would be class affiliation: working class members are more likely to drink beer than wine and are more probable to eat less healthy food and have less healthy jobs. By defining the groups solely by choice of beverage, the researchers have unwillingly split by the confounding variable of class affiliation.

In this case, the problem is more subtle & interesting than the usual correlation!=causation issues, and I owe it to the psychologist Paul Meehl: matching methods have the problem that when you match groups on some observed variables but different outcomes, you're guaranteeing they will be very weird in unknown ways.

His example was kids, schooling, and IQ; to 'control for' IQ you might try to compare kids with bad grades & high IQ with some other kids with good grades & high IQ, in order to find the effect of schooling on some outcome like crime. Except... kids with high IQ should have good grades; so why does that first group have bad grades? Why are they weird like that? What's going on? Do they have some sort of severe home problem? Emerging mental health issues? Systematic personality differences? You may well find that having 'controlled for' IQ, the kids with bad grades are more likely to be jailed and conclude that schooling helps prevent crime... but what have you really found?

In this case, what the researchers have found is that twins are powerfully locked together in terms of health outcomes, body types, diets, and exercise tolerance/preference, so locked together that they have to scour millions of people to turn up a handful of twin pairs. These twin pairs are 'matched' on genetics/upbringing (IQ), but not matched on exercise (grades), and so one gets health outcomes (jailing) 'controlling for' genetics/upbringing... so one concludes exercise helps prevent worse health... but what have you really found?

(At least, I think this is what Paul Meehl was talking about in http://www.mcps.umn.edu/assets/pdf/4.13_Meehl.pdf "Nuisance Variables and the Ex Post Facto Design", 1969, in _Minnesota studies in the philosophy of science: Vol. IV. Analyses of theories and methods of physics and psychology_ - that paper is unusually opaque for Meehl, and I'm not sure I understand it.)

"The dissimilarities in their exercise routines had mostly begun within the past three years, according to their questionnaires."

Seeing a significant change in three years is great. Imagine what it'll look like 10 years from now

Speaking of diet, I'm wondering what your thoughts are on macronutrient composition (e.g., low-carbohydrate) diets? I didn't find anything in your essays, just this series of tweets.[0]

[0] https://twitter.com/gwern/status/456548028745129984

>(Less than 1 on BMI? 2kg of weight? I fluctuate more than that on a weekly basis...)

If you weigh yourself at a controlled stage/time in your daily routine and fluctuate 2kg or more on a weekly basis, maybe you're doing something wrong.

Anecdata: I took up weight-lifting and running together and changed my diet. Presto changeo, over the past couple of years of doing this consistently I've lost about 12kg (EDIT: actually, no, that was in the single last February-February alone) while gaining lean muscle mass and gaining a few centimeters of height (that last part was sheer luck, of course). Oh, and my mood improved quite consistently, too.

Turned out being fat was mostly bad habits, with the problem being that I didn't think of them as alterable habits because I hadn't conceived of alternatives.

There are several sayings in fitness circles that are similar to this: "You can't outrun your fork." Diet is the primary reason people are obese, not lack of exercise.

I was about to post "can't outrun a fork" too. Hour of running five days a week burns only enough calories two account for two super-sized Big Mac meals.

The counterpoint would be that exercise changes your body in all kinds of ways, which might have a bigger overall effect.

Not sure if it's true, but it's at least as scientific as "can't outrun your fork".

I haven't seen much evidence that exercise Can significantly change your metabolic rate.

I haven't bothered to check if there's scientific literature that proves ingesting calories is easier than burning them, but I think you'd struggle to find a scientist or sane person who doesn't feel that generally it's orders of magnitude easier.

Assuming you are talking about 500kcal per exercise session: This results a total of 130000kcal per year, or the nutritional value of more than 16kg of fat. So how is your example an argument against outrunning your fork? If you enjoy both eating and exercise, this may as well be the way of least resistance.

Because your body will signal that it needs more calories, and then you'll eat more calories.

Yeah I've recently started counting calories without changing the amount I exercise and I'm finding it's not that hard to lose a little weight just by keeping the portion sizes down.

"You can't outrun your fork."

I can.

People who can't outrun their forks aren't running hard enough.

Hunting down a lot of calories and eating them is part of being a mammal. I bicycled across North America one fall on a mostly mountain route. Four times a day I ate full meals and added as many Dove bars as I could without being sick in between. I biked up and down hills with 30kg of gear all day. And I lost a lot of weight.

Sure it hurts to work that hard, but it's supposed to hurt a little. Being comfortable all the time isn't part of being a mammal.

It feels disingenuous to suggest the barrier to obese people not being obese is that they're not taking quarter-year bike trips across North America.

I look forward to your enlightening anecdote based debunking of other sayings, such as that time you made a great broth despite too many cooks, or that time a gathering didn't get merrier with more people.

I'm a little too busy to entertain you with vignettes right now; I need to supervise this pot closely until it boils.

I occasionally joke that I know the perfect diet: my wife and I went on a hard two week hike a few years ago. We both lost 10-15 lbs and neither of us had the weight to lose at the start. I make sure people know I am joking when I recommend it though: it would be a very bad idea for most people to try it, and that probably includes me.

By the way, if anyone wants to attempt an unsupported hike of the John Muir trail, send me an email and I will be happy to give you some advice.

I am a very, very picky eater.

The first time I came to China, I didn't know how to find palatable food. (Yes, this was something I worried about beforehand.) Over the course of three months, I lost 15 pounds.

I had been variously described, before that point, as "thin", "stick-thin", and "skeletal".

When my dad picked me up at the airport afterwards, he was shocked. He said I looked like I was literally starving to death, which in a limited sense was true. I spent a lot of those three months visualizing pizza.

Was it the perfect diet? I'm sure the approach (just refuse to eat what's available) works. But it was a truly awful experience. Not recommended to anyone.

Most people are fat because of diet, and just exercising a bit is not going to help them as much as they might hope.

Your anecdotate is just a silly brag that adds nothing to the discussion.

It's not so simple. State of fitness/wellness can be interpreted as a mindset.

With this interpretation, exercising and eating adequately are very strictly related, I'd say they have the same root.

But also from a more physical point of view, nutrition is also a function of exercise, since there are many needs of the body which are function of it. So it goes both ways, I wouldn't say one of them is primary.